Operation and configuration reference for Edge Microgateway

You're viewing Apigee Edge documentation.
Go to the Apigee X documentation. info

Edge Microgateway v. 3.3.x

This topic discusses how to manage and configure Edge Microgateway.

Upgrading Edge Microgateway if you have an internet connection

This section explains how to upgrade an existing installation of Edge Microgateway. If you are operating without an internet connection, see Can I install Edge Microgateway without an internet connection?.

Apigee recommends you test your existing configuration with the new version before upgrading your production environment.

1. Execute the following npm command to upgrade to the latest version of Edge Microgateway:

   npm upgrade edgemicro -g

   To install a specific version of Edge Microgateway, you need to specify the version number in the install command. For example, to install to version 3.2.3, use the following command:

   npm install edgemicro@3.2.3 -g

2. Check the version number. For example, if you installed version 3.2.3:

   edgemicro --version
   current nodejs version is v12.5.0
   current edgemicro version is 3.2.3
       

3. Finally, upgrade to the latest version of the edgemicro-auth proxy:

   edgemicro upgradeauth -o $ORG -e $ENV -u $USERNAME

Making configuration changes

The configuration files you need to know about include:

• Default system configuration file
• Default config file for a newly initialized Edge Microgateway instance
• Dynamic configuration file for running instances

This section discusses these files and what you need to know about changing them.

Default system configuration file

When you install Edge Microgateway, a default system configuration file is placed here:

prefix/lib/node_modules/edgemicro/config/default.yaml

Where prefix is the npm prefix directory. See Where is Edge Microgateway installed if you can't locate this directory.

If you change the system config file, you must reinitialize, reconfigure, and restart Edge Microgateway:

edgemicro init
edgemicro configure [params]
edgemicro start [params]

Default config file for newly initialized Edge Microgateway instances

When you run edgemicro init, the system config file (described above), default.yaml, is placed in the ~/.edgemicro directory.

If you change the config file in ~/.edgemicro, you must reconfigure and restart Edge Microgateway:

edgemicro stop
edgemicro configure [params]
edgemicro start [params]

Dynamic configuration file for running instances

When you run edgemicro configure [params], a dynamic configuration file is created in ~/.edgemicro. The file is named according to this pattern: org-env-config.yaml, where org and env are your Apigee Edge organization and environment names. You can use this file to make configuration changes, and then reload them with zero-downtime. For example, if you add and configure a plugin, you can reload the configuration without incurring any downtime, as explained below.

If Edge Microgateway is running (zero-downtime option):

1. Reload the Edge Microgateway configuration:

   edgemicro reload -o $ORG -e $ENV -k $KEY -s $SECRET

   Where:

   • $ORG is your Edge organization name (you must be an organization administrator).
   • $ENV is an environment in your org (such as "test" or "prod").
   • $KEY is the key returned previously by the configure command.
   • $SECRET is the key returned previously by the configure command.

   For example

   edgemicro reload -o docs -e test -k 701e70ee718ce6dc188...78b6181d000723 \
     -s 05c14356e42ed1...4e34ab0cc824

If Edge Microgateway is stopped:

1. Restart Edge Microgateway:

   edgemicro start -o $ORG -e $ENV -k $KEY -s $SECRET

   Where:

   • $ORG is your Edge organization name (you must be an organization administrator).
   • $ENV is an environment in your organization (such as "test" or "prod").
   • $KEY is the key returned previously by the configure command.
   • $SECRET is the key returned previously by the configure command.

   For example:

   edgemicro start -o docs -e test -k 701e70ee718ce...b6181d000723 \
     -s 05c1435...e34ab0cc824

Here is an example config file. For details about configuration file settings, see Edge Microgateway configuration reference.

edge_config:
  bootstrap: >-
    https://edgemicroservices-us-east-1.apigee.net/edgemicro/bootstrap/organization/docs/environment/test
  jwt_public_key: 'https://docs-test.apigee.net/edgemicro-auth/publicKey'
  managementUri: 'https://api.enterprise.apigee.com'
  vaultName: microgateway
  authUri: 'https://%s-%s.apigee.net/edgemicro-auth'
  baseUri: >-
    https://edgemicroservices.apigee.net/edgemicro/%s/organization/%s/environment/%s
  bootstrapMessage: Please copy the following property to the edge micro agent config
  keySecretMessage: The following credentials are required to start edge micro
  products: 'https://docs-test.apigee.net/edgemicro-auth/products'
edgemicro:
  port: 8000
  max_connections: 1000
  max_connections_hard: 5000
  config_change_poll_interval: 600
  logging:
    level: error
    dir: /var/tmp
    stats_log_interval: 60
    rotate_interval: 24
  plugins:
    sequence:
      - oauth
headers:
  x-forwarded-for: true
  x-forwarded-host: true
  x-request-id: true
  x-response-time: true
  via: true
oauth:
  allowNoAuthorization: false
  allowInvalidAuthorization: false
  verify_api_key_url: 'https://docs-test.apigee.net/edgemicro-auth/verifyApiKey'
analytics:
  uri: >-
    https://edgemicroservices-us-east-1.apigee.net/edgemicro/axpublisher/organization/docs/environment/test

Setting environment variables

The command-line interface commands that require values for your Edge organization and environment, and the key and secret needed for starting Edge Microgateway can be stored in these environment variables:

• EDGEMICRO_ORG
• EDGEMICRO_ENV
• EDGEMICRO_KEY
• EDGEMICRO_SECRET

Setting these variables is optional. If you set them, you do not have to specify their values when you use the Command-Line Interface (CLI) to configure and start Edge Microgateway.

Configuring SSL on the Edge Microgateway server

Watch the following videos to learn about configuring TLS in Apigee Edge Microgateway:

You can configure the Microgateway server to use SSL. For example, with SSL configured, you can call APIs through Edge Microgateway with the "https" protocol, like this:

https://localhost:8000/myapi

To configure SSL on the Microgateway server, follow these steps:

1. Generate or obtain an SSL certificate and key using the openssl utility or whichever method you prefer.
2. Add the edgemicro:ssl attribute to the Edge Microgateway configuration file. For a complete list of options, see the table below. For example:
   

   edgemicro:
     ssl:
      key: <absolute path to the SSL key file>
      cert: <absolute path to the SSL cert file>
      passphrase: admin123 #option added in v2.2.2
      rejectUnauthorized: true #option added in v2.2.2
      requestCert: true

3. Restart Edge Microgateway. Follow the steps outlined in Making configuration changes depending on which configuration file you edited: the default file or the runtime config file.

Here's an example of the edgemicro section of the config file, with SSL configured:

edgemicro:
  port: 8000
  max_connections: 1000
  max_connections_hard: 5000
  logging:
    level: error
    dir: /var/tmp
    stats_log_interval: 60
    rotate_interval: 24
  plugins:
    sequence:
      - oauth
  ssl:
    key: /MyHome/SSL/em-ssl-keys/server.key
    cert: /MyHome/SSL/em-ssl-keys/server.crt
    passphrase: admin123 #option added in v2.2.2
    rejectUnauthorized: true #option added in v2.2.2

Here is a list of all of the supported server options:

Using client SSL/TLS options

You can configure Edge Microgateway to be a TLS or SSL client when connecting to target endpoints. In the Microgateway configuration file, use the targets element to set SSL/TLS options. Note that you can specify multiple specific targets. A multi-target example is included below.

This example provides settings that will be applied to all hosts:

edgemicro:
...
targets:
  ssl:
    client:
      key: /Users/jdoe/nodecellar/twowayssl/ssl/client.key
      cert: /Users/jdoe/nodecellar/twowayssl/ssl/ca.crt
      passphrase: admin123
      rejectUnauthorized: true

In this example, the settings are applied only to the specified host:

edgemicro:
...
targets:
  - host: 'myserver.example.com'
    ssl:
      client:
        key: /Users/myname/twowayssl/ssl/client.key
        cert: /Users/myname/twowayssl/ssl/ca.crt
        passphrase: admin123
        rejectUnauthorized: true

Here is an example for TLS:

edgemicro:
...
targets:
  - host: 'myserver.example.com'
    tls:
      client:
        pfx: /Users/myname/twowayssl/ssl/client.pfx
        passphrase: admin123
        rejectUnauthorized: true

In the case where you want to apply TLS/SSL settings to multiple specific targets, you must specify the first host in the configuration as "empty", which enables universal requests, and then specify specific hosts in any order. In this example, the settings are applied to multiple specific hosts:

targets:
 - host:   ## Note that this value must be "empty"
   ssl:
     client:
       key: /Users/myname/twowayssl/ssl/client.key
       cert: /Users/myname/twowayssl/ssl/ca.crt
       passphrase: admin123
       rejectUnauthorized: true
 - host: 'myserver1.example.com'
   ssl:
     client:
       key: /Users/myname/twowayssl/ssl/client.key
       cert: /Users/myname/twowayssl/ssl/ca.crt
       rejectUnauthorized: true
 - host: 'myserver2.example.com'
   ssl:
     client:
       key: /Users/myname/twowayssl/ssl/client.key
       cert: /Users/myname/twowayssl/ssl/ca.crt
       rejectUnauthorized: true

Here is a list of all of the supported client options:

Customizing the edgemicro-auth proxy

By default, Edge Microgateway uses a proxy deployed on Apigee Edge for OAuth2 authentication. This proxy is deployed when you initially run edgemicro configure. You can change the default configuration of this proxy to add support for custom claims to a JSON Web Token (JWT), configure token expiration, and generate refresh tokens. For details, see the edgemicro-auth page in GitHub.

Using a custom auth service

By default, Edge Microgateway uses a proxy deployed on Apigee Edge for OAuth2 authentication. This proxy is deployed when you initially run edgemicro configure. By default, this proxy's URL is specified in the Edge Microgateway config file as follows:

authUri: https://myorg-myenv.apigee.net/edgemicro-auth

If you want to use your own custom service to handle authentication, change the authUri value in the config file to point to your service. For example, you may have a service that uses LDAP to verify identity.

Managing log files

Edge Microgateway logs information about each request and response. Log files provide useful information for debugging and troubleshooting.

Where log files are stored

By default, log files are stored in /var/tmp.

How to change the default log file directory

The directory where log files are stored is specified in the Edge Microgateway configuration file. See also Making configuration changes.

edgemicro:
  home: ../gateway
  port: 8000
  max_connections: -1
  max_connections_hard: -1
  logging:
    level: info
    dir: /var/tmp
    stats_log_interval: 60
    rotate_interval: 24

Change the dir value to specify a different log file directory.

Send logs to the console

You can configure logging so that log information is sent to standard output instead of to a log file. Set the to_console flag to true as follows:

edgemicro:
  logging:
    to_console: true

With this setting, logs will be sent to standard out. Currently, you cannot send logs to both stdout and to a log file.

How to set the logging level

You specify the log level to use in the edgemicro configuration. For a complete list of log levels and their descriptions, see edgemicro attributes.

For example, the following configuration sets the logging level to debug:

edgemicro:
  home: ../gateway
  port: 8000
  max_connections: -1
  max_connections_hard: -1
  logging:
    level: debug
    dir: /var/tmp
    stats_log_interval: 60
    rotate_interval: 24

How to change log intervals

You can configure these intervals in the Edge Microgateway config file. See also Making configuration changes.

The configurable attributes are:

• stats_log_interval: (default: 60) Interval, in seconds, when the stats record is written to the API log file.
• rotate_interval: (default: 24) Interval, in hours, when log files are rotated. For example:

edgemicro:
  home: ../gateway
  port: 8000
  max_connections: -1
  max_connections_hard: -1
  logging:
    level: info
    dir: /var/tmp
    stats_log_interval: 60
    rotate_interval: 24

How to relax strict log file permissions

By default, Edge Microgateway generates the application log file (api-log.log) with the file permission level set to 0600. This permission level does not allow external applications or users to read the log file. To relax this strict permission level, set logging:disableStrictLogFile to true. When this attribute is true, the log file is created with the file permission set to 0755. If false or if the attribute is not provided, permission defaults to 0600.

Added in v3.2.3.

For example:

edgemicro:
 logging:
   disableStrictLogFile: true

Good log file maintenance practices

As log file data accumulates over time, Apigee recommends that you adopt the following practices:

• Because log files can become quite large, be sure that the log file directory has sufficient space. See the following sections Where log files are stored and How to change the default log file directory.
• Either delete or move log files to a separate archive directory at least once a week.
• If your policy is to delete logs, you can use the CLI command edgemicro log -c to remove (clean) older logs.

Log file naming convention

Each Edge Microgateway instance produces a log file with a .log extension. The naming convention for log files is as follows:

edgemicro-HOST_NAME-INSTANCE_ID-api.log

For example:

edgemicro-mymachine-local-MTQzNTgNDMxODAyMQ-api.log

About log file contents

Added in: v2.3.3

By default, the logging service omits the JSON of downloaded proxies, products, and the JSON Web Token (JWT). If you wish to output these objects to the console, set the command-line flag DEBUG=* when you start Edge Microgateway. For example:

DEBUG=* edgemicro start -o docs -e test -k abc123 -s xyz456

Contents of the "api" log file

The "api" log file contains detailed information about the flow of requests and responses through Edge Microgateway. The "api" log files are named like this:

edgemicro-mymachine-local-MTQzNjIxOTk0NzY0Nw-api.log

For each request made to Edge Microgateway, four events are captured in the "api" log file:

• Incoming request from the client
• Outgoing request made to the target
• Incoming response from the target
• Outgoing response to the client

Each of these separate entries is represented in a shorthand notation to help make the log files more compact. Here are four sample entries representing each of the four events. In the log file, they look like this (the line numbers are only for reference in the doc, they don't appear in the log file).

(1) 1436403888651 info req m=GET, u=/, h=localhost:8000, r=::1:59715, i=0
(2) 1436403888665 info treq m=GET, u=/, h=127.0.0.18080, i=0
(3) 1436403888672 info tres s=200, d=7, i=0
(4) 1436403888676 info res s=200, d=11, i=0

Let's look at them one by one:

1. Sample of incoming request from client:

1436403888651 info req m=GET, u=/, h=localhost:8000, r=::1:59715, i=0

• 1436403888651 - Unix date stamp
• info - The logging level. This value depends on context of the transaction and the logging level set in the edgemicro configuration. See How to set the logging level. For stats records, the level is set to stats. Stats records are reported at a regular interval set with the stats_log_interval configuration. See also How to change log intervals.
• req - Identifies the event. In this case, request from the client.
• m - The HTTP verb used in the request.
• u - The part of the URL following the basepath.
• h - The host and port number where Edge Microgateway is listening.
• r - The remote host and port where the client request originated.
• i - The request ID. All four event entries will share this ID. Each request is assigned a unique request ID. Correlating log records by request ID can give valuable insight into the target’s latency.
• d - The duration in milliseconds since the request was received by Edge Microgateway. In the example above, the target’s response for request 0 was received after 7 milliseconds (line 3), and the response was sent to the client after an additional 4 milliseconds (line 4). In other words, the total request latency was 11 milliseconds, out of which 7 milliseconds were taken by the target and 4 milliseconds by Edge Microgateway itself.

2. Sample of outgoing request made to the target:

1436403888665 info treq m=GET, u=/, h=127.0.0.1:8080, i=0

• 1436403888651 - Unix date stamp
• info - The logging level. This value depends on context of the transaction and the logging level set in the edgemicro configuration. See How to set the logging level. For stats records, the level is set to stats. Stats records are reported at a regular interval set with the stats_log_interval configuration. See also How to change log intervals.
• treq - Identifies the event. In this case, target request.
• m - The HTTP verb used in the target request.
• u - The part of the URL following the basepath.
• h - The host and port number of the backend target.
• i - The ID of the log entry. All four event entries will share this ID.

3. Sample of incoming response from the target

1436403888672 info tres s=200, d=7, i=0

1436403888651 - Unix date stamp

• info - The logging level. This value depends on context of the transaction and the logging level set in the edgemicro configuration. See How to set the logging level. For stats records, the level is set to stats. Stats records are reported at a regular interval set with the stats_log_interval configuration. See also How to change log intervals.
• tres - Identifies the event. In this case, target response.
• s - The HTTP response status.
• d - The duration in milliseconds. The time taken for the API call by the target.
• i - The ID of the log entry. All four event entries will share this ID.

4. Sample of outgoing response to the client

1436403888676 info res s=200, d=11, i=0

1436403888651 - Unix date stamp

• info - The logging level. This value depends on context of the transaction and the logging level set in the edgemicro configuration. See How to set the logging level. For stats records, the level is set to stats. Stats records are reported at a regular interval set with the stats_log_interval configuration. See also How to change log intervals.
• res - Identifies the event. In this case, response to the client.
• s - The HTTP response status.
• d - The duration in milliseconds. This is the total time taken by the API call, including the time taken by the target API and the time take by Edge Microgateway itself.
• i - The ID of the log entry. All four event entries will share this ID.

Log file schedule

Log files are rotated on the interval specified by the rotate_interval configuration attribute. Entries will continue being added to the same log file until the rotation interval expires. However, each time Edge Microgateway is restarted it receives a new UID and creates a new set of log files with this UID. See also Good log file maintenance practices.

Error messages

Some log entries will contain error messages. To help identify where and why the errors occur, see the Edge Microgateway error reference.

Edge Microgateway configuration reference

Location of the configuration file

The configurations attributes described in this section are located in the Edge Microgateway configuration file. See also Making configuration changes.

edge_config attributes

These settings are used to configure interaction between the Edge Microgateway instance and Apigee Edge.

• bootstrap: (default: none) A URL that points to an Edge Microgateway-specific service running on Apigee Edge. Edge Microgateway uses this service to communicate with Apigee Edge. This URL is returned when you execute the command to generate the public/private key pair: edgemicro genkeys. See the Setting up and configuring Edge Microgateway for details.
• jwt_public_key: (default: none) A URL that points to the Edge Microgateway proxy that is deployed on Apigee Edge. This proxy serves as an authentication endpoint for issuing signed access tokens to clients. This URL is returned when you execute the command to deploy the proxy: edgemicro configure. See the Setting up and configuring Edge Microgateway for details.
• quotaUri: Set this config property if you want to manage quotas through the edgemicro-auth proxy that is deployed to your org. If this property is not set, the quota endpoint defaults to the internal Edge Microgateway endpoint.

  edge_config:
    quotaUri: https://your_org-your_env.apigee.net/edgemicro-auth
  

edgemicro attributes

These settings configure the Edge Microgateway process.

• port: (default: 8000) The port number on which the Edge Microgateway process listens.
• max_connections: (default: -1) Specifies the maximum number of simultaneous incoming connections that Edge Microgateway can receive. If this number is exceeded, the following status is returned:
  
  res.statusCode = 429; // Too many requests
  
• max_connections_hard: (default: -1) The maximum number of simultaneous requests that Edge Microgateway can receive before shutting down the connection. This setting is intended to thwart denial of service attacks. Typically, set it to a number larger than max_connections.
• logging:
  • level: (default: error)
    • info - (Recommended) Logs all requests and responses that flow through an Edge Microgateway instance.
    • warn - Logs warning messages only.
    • error - Logs error messages only.
    • debug - Logs debug messages along with info, warn, and error messages.
    • trace - Logs trace information for errors along with info, warn, and error messages.
    • none - Do not create a log file.
  • dir: (default: /var/tmp) The directory where log files are stored.
  • stats_log_interval: (default: 60) Interval, in seconds, when the stats record is written to the api log file.
  • rotate_interval: (default: 24) Interval, in hours, when log files are rotated.

• plugins: Plugins add functionality to Edge Microgateway. For details about developing plugins, see Develop custom plugins.

• dir: A relative path from ./gateway directory to the ./plugins directory, or an absolute path.
• sequence: A list of plugin modules to add to your Edge Microgateway instance. The modules will execute in the order they are specified here.

• debug: Adds remote debugging to the Edge Microgateway process.
  • port: The port number to listen on. For example, set your IDE debugger to listen on this port.
  • args: Arguments to the debug process. For example: args --nolazy
    
• config_change_poll_interval: (default: 600 seconds) Edge Microgateway loads a new configuration periodically and executes a reload if anything changed. The polling picks up any changes made on Edge (changes to products, microgateway-aware proxies, etc) as well as changes made to the local config file.
  
• disable_config_poll_interval: (default: false) Set to true to turn off automatic change polling.
• request_timeout: Sets a timeout for target requests. The timeout is set in seconds. If a timeout occurs, Edge Microgateway responds with a 504 status code. (Added v2.4.x)
• keep_alive_timeout: This property enables you to set the Edge Microgateway timeout (in milliseconds). (Default: 5 seconds) (Added v3.0.6)
• headers_timeout: This attribute limits the amount of time (in milliseconds) the HTTP parser will wait to receive the complete HTTP headers.

  For example:

  edgemicro:
    keep_alive_timeout: 6000
    headers_timeout: 12000

  Internally, the parameter sets the Node.js Server.headersTimeout attribute on requests. (Default: 5 seconds more than the time set with edgemicro.keep_alive_timeout. This default setting prevents load balancers or proxies from erroneously dropping the connection.) (Added v3.1.1)

• noRuleMatchAction: (String) The action to take (allow or deny access) if the match rule specified in the accesscontrol plugin is not resolved (is unmatched). Valid values: ALLOW or DENY Default: ALLOW (Added: v3.1.7)
• enableAnalytics: (default: true) Set the attribute to false to prevent the analytics plugin from being loaded. In this case, no calls to Apigee Edge analytics will be made. If set to true, or when this attribute is not provided, analytics plugin will work as usual. See edgemicro attributes for details. (Added v3.1.8).

  Example:

  edgemicro
    enableAnalytics=false|true

• on_target_response_abort: This attribute lets you control how Edge Microgateway behaves if the connection between the client (Edge Microgateway) and the target server closes prematurely.

  Value	Description
  Default	If on_target_response_abort is unspecified, then the default behavior is to truncate the response without showing an error. In log files, a warning message is shown with targetResponse aborted and a 502 response code.
  appendErrorToClientResponseBody	The custom error TargetResponseAborted is returned to the client. In log files, a warning message is shown with targetResponse aborted and a 502 response code. In addition, the error TargetResponseAborted is logged with the message Target response ended prematurely.
  abortClientRequest	Edge Microgateway aborts the request and a warning is written to the log files: TargetResponseAborted with the 502 request status code.

Example:

edgemicro:
 on_target_response_abort: appendErrorToClientResponseBody | abortClientRequest

headers attributes

These settings configure how certain HTTP headers are treated.

• x-forwarded-for: (default: true) Set to false to prevent x-forwarded-for headers to be passed to the target. Note that if an x-forwarded-for header is in the request, its value will be set to the client-ip value in Edge Analytics.
• x-forwarded-host: (default: true) Set to false to prevent x-forwarded-host headers to be passed to the target.
• x-request-id: (default: true) Set to false to prevent x-request-id headers to be passed to the target.
• x-response-time: (default: true) Set to false to prevent x-response-time headers to be passed to the target.
• via: (default: true) Set to false to prevent via headers to be passed to the target.

oauth attributes

These settings configure how client authentication is enforced by Edge Microgateway.

• allowNoAuthorization: (default: false) If set to true, API calls are allowed to pass through Edge Microgateway without any Authorization header at all. Set this to false to require an Authorization header (default).
• allowInvalidAuthorization: (default: false) If set to true, API calls are allowed to pass if the token passed in the Authorization header is invalid or expired. Set this to false to require valid tokens (default).
• authorization-header: (default: Authorization: Bearer) The header used to send the access token to Edge Microgateway. You may wish to change the default in cases where the target needs to use the Authorization header for some other purpose.
• api-key-header: (default: x-api-key) The name of the header or query parameter used to pass an API key to Edge Microgateway. See also Using an API key.
• keep-authorization-header: (default: false) If set to true, the Authorization header sent in the request is passed on to the target (it is preserved).
• allowOAuthOnly -- If set to true, every API must carry an Authorization header with a Bearer Access Token. Allows you to permit only the OAuth security model (while maintaining backward compatibility). (Added 2.4.x)
• allowAPIKeyOnly -- If set to true, every API must carry an x-api-key header (or a custom location) with an API Key.Allows you to permit only the API key security model (while maintaining backward compatibility). (Added 2.4.x)
• gracePeriod -- This parameter helps prevent errors caused by slight discrepancies between your system clock and the Not Before (nbf) or Issued At (iat) times specified in the JWT authorization token. Set this parameter to the number of seconds to allow for such discrepancies. (Added 2.5.7)

Plugin-specific attributes

See Using plugins for details on configurable attributes for each plugin.

Filtering proxies

You can filter which microgateway-aware proxies an Edge Microgateway instance will process. When Edge Microgateway starts, it downloads all of the microgateway-aware proxies in the organization it's associated with. Use the following configuration to limit which proxies the microgateway will process. For example, this configuration limits the proxies the microgateway will process to three: edgemicro_proxy-1, edgemicro_proxy-2, and edgemicro_proxy-3:

edgemicro:
  proxies:
  - edgemicro_proxy-1
  - edgemicro_proxy-2
  - edgemicro_proxy-3

Filtering products by name

Use the following configuration to limit the number of API products that Edge Microgateway downloads and processes. To filter downloaded products, add the productnamefilter query parameter to the /products API listed in the Edge Microgateway *.config.yaml file. For example:

edge_config:
  bootstrap: >-
    https://edgemicroservices.apigee.net/edgemicro/bootstrap/organization/willwitman/environment/test
  jwt_public_key: 'https://myorg-test.apigee.net/edgemicro-auth/publicKey'
  managementUri: 'https://api.enterprise.apigee.com'
  vaultName: microgateway
  authUri: 'https://%s-%s.apigee.net/edgemicro-auth'
  baseUri: >-
    https://edgemicroservices.apigee.net/edgemicro/%s/organization/%s/environment/%s
  bootstrapMessage: Please copy the following property to the edge micro agent config
  keySecretMessage: The following credentials are required to start edge micro
  products: 'https://myorg-test.apigee.net/edgemicro-auth/products?productnamefilter=%5E%5BEe%5Ddgemicro.%2A%24'

Note that the value of the query parameter must be specified in regular expression format and be URL encoded. For example, the regex ^[Ee]dgemicro.*$ catches names such as: "edgemicro-test-1" , "edgemicro_demo" and "Edgemicro_New_Demo". The URL encoded value, suitable for use in the query parameter, is: %5E%5BEe%5Ddgemicro.%2A%24.

The following debug output shows that only the filtered products were downloaded:

...
2020-05-27T03:13:50.087Z [76060] [microgateway-config network] products download from https://gsc-demo-prod.apigee.net/edgemicro-auth/products?productnamefilter=%5E%5BEe%5Ddgemicro.%2A%24 returned 200 OK
...
....
....
{
   "apiProduct":[
      {
         "apiResources":[

         ],
         "approvalType":"auto",
         "attributes":[
            {
               "name":"access",
               "value":"public"
            }
         ],
         "createdAt":1590549037549,
         "createdBy":"k***@g********m",
         "displayName":"test upper case in name",
         "environments":[
            "prod",
            "test"
         ],
         "lastModifiedAt":1590549037549,
         "lastModifiedBy":"k***@g********m",
         "name":"Edgemicro_New_Demo",
         "proxies":[
            "catchall"
         ],
         "quota":"null",
         "quotaInterval":"null",
         "quotaTimeUnit":"null",
         "scopes":[

         ]
      },
      {
         "apiResources":[

         ],
         "approvalType":"auto",
         "attributes":[
            {
               "name":"access",
               "value":"public"
            }
         ],
         "createdAt":1590548328998,
         "createdBy":"k***@g********m",
         "displayName":"edgemicro test 1",
         "environments":[
            "prod",
            "test"
         ],
         "lastModifiedAt":1590548328998,
         "lastModifiedBy":"k***@g********m",
         "name":"edgemicro-test-1",
         "proxies":[
            "Lets-Encrypt-Validation-DoNotDelete"
         ],
         "quota":"null",
         "quotaInterval":"null",
         "quotaTimeUnit":"null",
         "scopes":[

         ]
      },
      {
         "apiResources":[
            "/",
            "/**"
         ],
         "approvalType":"auto",
         "attributes":[
            {
               "name":"access",
               "value":"public"
            }
         ],
         "createdAt":1558182193472,
         "createdBy":"m*********@g********m",
         "displayName":"Edge microgateway demo product",
         "environments":[
            "prod",
            "test"
         ],
         "lastModifiedAt":1569077897465,
         "lastModifiedBy":"m*********@g********m",
         "name":"edgemicro_demo",
         "proxies":[
            "edgemicro-auth",
            "edgemicro_hello"
         ],
         "quota":"600",
         "quotaInterval":"1",
         "quotaTimeUnit":"minute",
         "scopes":[

         ]
      }
   ]
}

Filtering products by custom attributes

To filter products based on custom attributes:

1. In the Edge UI, select the edgemicro_auth proxy in the organization/environment where you configured Edge Microgateway.
2. In the Develop tap, open the JavaCallout policy in the editor.
3. Add a custom attribute with the key products.filter.attributes with a comma-separated list of attribute names. Only products that contain any of the custom attribute names will be returned to Edge Microgateway.
4. You can optionally disable the check to see if the product is enabled for the current environment by setting the custom attribute products.filter.env.enable to false. (The default is true.)
5. (Private Cloud Only) If you are on Edge for Private Cloud, set the property org.noncps to true to pull products for non-CPS environments.

For example:

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<JavaCallout async="false" continueOnError="false" enabled="true" name="JavaCallout">
    <DisplayName>JavaCallout</DisplayName>
    <FaultRules/>
    <Properties>
        <Property name="products.filter.attributes">attrib.one, attrib.two</Property>
        <Property name="products.filter.env.enable">false</Property>
        <Property name="org.noncps">true</Property>
    </Properties>
    <ClassName>io.apigee.microgateway.javacallout.Callout</ClassName>
    <ResourceURL>java://micro-gateway-products-javacallout-2.0.0.jar</ResourceURL>
</JavaCallout>

Configuring analytics push frequency

Use these configuration parameters to control the frequency at which Edge Microgateway sends analytics data to Apigee:

• bufferSize (Optional): The maximum number of analytics records that the buffer can hold before beginning to drop the oldest records. Default: 10000
• batchSize (Optional): The maximum size of a batch of analytics records sent to Apigee. Default: 500
• flushInterval (Optional): The number of milliseconds between each flush of a batch of analytics records sent to Apigee. Default: 5000

For example:

analytics:
  bufferSize: 15000
  batchSize: 1000
  flushInterval: 6000

Masking analytics data

The following configuration prevents request path information from showing up in Edge analytics. Add the following to the microgateway configuration to mask the request URI and/or request path. Note that the URI consists of the hostname and path parts of the request.

analytics:
  mask_request_uri: 'string_to_mask'
  mask_request_path: 'string_to_mask'

Segregating API calls in Edge Analytics

You can configure the analytics plugin to segregate a specific API path so that it appears as a separate proxy in the Edge Analytics dashboards. For example, you can segregate a health check API in the dashboard to avoid confusing it with actual API proxy calls. In the Analytics dashboard, segregated proxies follow this naming pattern:

edgemicro_proxyname-health

The following image shows two segregated proxies in the Analytics dashboard: edgemicro_hello-health and edgemicro_mock-health:

Use these parameters to segregate relative and absolute paths in the Analytics dashboard as separate proxies:

• relativePath (Optional): Specifies a relative path to segregate in the Analytics dashboard. For example, if you specify /healthcheck, all API calls that contain the path /healthcheck will appear in the dashboard as edgemicro_proxyname-health. Note that this flag ignores the proxy basepath. To segregate based on a full path, including basepath, use the proxyPath flag.
• proxyPath (Optional): Specifies a full API proxy path, including the proxy basepath, to segregate in the analytics dashboard. For example, if you specify /mocktarget/healthcheck, where /mocktarget is the proxy basepath, all API calls with the path /mocktarget/healthcheck will appear in the dashboard as edgemicro_proxyname-health.

For example, in the following configuration any API path that contains /healthcheck will be segregated by the analytics plugin. This means, /foo/healthcheck and /foo/bar/healthcheck will be segregated as a separate proxy called edgemicro_proxyname-health in the analytics dashboard.

analytics:
  uri: >-
    https://xx/edgemicro/ax/org/docs/environment/test
  bufferSize: 100
  batchSize: 50
  flushInterval: 500
  relativePath: /healthcheck

In the following configuration any API with the proxy path /mocktarget/healthcheck will be will be segregated as a separate proxy called edgemicro_proxyname-health in the analytics dashboard.

analytics:
  uri: >-
    https://xx/edgemicro/ax/org/docs/environment/test
  bufferSize: 100
  batchSize: 50
  flushInterval: 500
  proxyPath: /mocktarget/healthcheck

Setting up Edge Microgateway behind a company firewall

Use an HTTP proxy for communication with Apigee Edge

Added in version 3.1.2.

To use an HTTP proxy for communication between Edge Microgateway and Apigee Edge, do the following:

1. Set the environment variables HTTP_PROXY, HTTPS_PROXY, and NO_PROXY. These variables control the hosts for each HTTP proxy that you wish to use for communication with Apigee Edge, or which hosts should not handle communication with Apigee Edge. For example:

   export HTTP_PROXY='http://localhost:3786'
   export HTTPS_PROXY='https://localhost:3786'
   export NO_PROXY='localhost,localhost:8080'

   Note that NO_PROXY can be a comma delimited list of domains that Edge Microgateway should not proxy to.

   For more information on these variables, see https://www.npmjs.com/package/request#controlling-proxy-behaviour-using-environment-variables

2. Restart Edge Microgateway.

Use an HTTP proxy for target communication

Added in version 3.1.2.

To use an HTTP proxy for communication between Edge Microgateway and backend targets, do the following:

1. Add the following configuration to the microgateway config file:

   edgemicro:
     proxy:
       tunnel: true | false
       url: proxy_url
       bypass: target_host # target hosts to bypass the proxy.
       enabled: true | false

   Where:

   • tunnel: (Optional) When true, Edge Microgateway uses the HTTP CONNECT method to tunnel HTTP requests over a single TCP connection. (The same is true if the environment variables, as mentioned below, for configuring the proxy are TLS enabled). Default: false
   • url: The HTTP proxy URL.
   • bypass: (Optional) Specifies one or more comma-separated target host URLs that should bypass the HTTP proxy. If this property is not set, then use the NO_PROXY environment variable to specify which target URLs to bypass.
   • enabled: If true and proxy.url is set, use the proxy.url value for the HTTP proxy. If true and proxy.url is not set, use the proxies specified in the HTTP proxy environment variables HTTP_PROXY and HTTPS_PROXY, as described in Use an HTTP proxy for communication with Apigee Edge.

   For example:

   edgemicro:
     proxy:
       tunnel: true
       url: 'http://localhost:3786'
       bypass: 'localhost','localhost:8080' # target hosts to bypass the proxy.
       enabled: true

2. Restart Edge Microgateway.

Using wildcards in Microgateway-aware proxies

You can use one or more "*" wildcards in the base path of an edgemicro_* (Microgateway-aware) proxy. For example, a base path of /team/*/members allows clients to call https://[host]/team/blue/members and https://[host]/team/green/members without you needing to create new API proxies to support new teams. Note that /**/ is not supported.

Important: Apigee does NOT support using a wildcard "*" as the first element of a base path. For example, this is NOT supported: /*/ search.

Rotating JWT keys

At some time after you initially generate a JWT, you might need to change the public/private key pair stored in the Edge encrypted KVM. This process of generating a new key pair is called key rotation.

How Edge Microgateway uses JWTs

JSON Web Token (JWT) is a token standard described in RFC7519. JWT provides a way to sign a set of claims, which can be verified reliably by the recipient of the JWT.

You can generate a JWT using the CLI and use it in the Authorization header of API calls instead of an API key. For example:

curl -i http://localhost:8000/hello -H "Authorization: Bearer eyJhbGciOiJ..dXDefZEA"

For information on generating JWTs with the CLI, see Generate a token.

What is key rotation?

At some time after you initially generate a JWT, you might need to change the public/private key pair stored in the Edge encrypted KVM. This process of generating a new key pair is called key rotation. When you rotate keys, a new private/public key pair is generated and stored in the "microgateway" KVM in your Apigee Edge organization/environment. In addition, the old public key is retained along with its original key ID value.

To generate a JWT, Edge uses information stored in the encrypted KVM. A KVM called microgatewaywas created and populated with keys when you initially set up (configured) Edge Microgateway. The keys in the KVM are used to sign and encrypt a JWT.

The KVM keys include:

• private_key - The latest (most recently created) RSA private key used to sign JWTs.

• public_key - The latest (most recently created) certificate used to verify JWTs signed with the private_key.

• private_key_kid - The latest (most recently created) private key ID. This key ID is associated with the private_key value and is used to support key rotation.

• public_key1_kid - The latest (most recently created) public key ID. This key is associated with the public_key1 value and is used to support key rotation. This value is the same as the private key kid.

• public_key1 - The latest (most recently created) public key.

When you perform key rotation, existing key values are replaced in the map and new keys are added to retain the old public keys. For example:

• public_key2_kid - The old public key ID. This key is associated with the public_key2 value and is used to support key rotation.

• public_key2 - The old public key.

JWTs presented for verification will be verified using the new public key. If verification fails, then the old public key will be used, until the JWT expires (after token_expiry* interval, default 30 mins). In this way, you can "rotate" keys without immediately disrupting API traffic.

How to do key rotation

This section explains how to perform a key rotation.

1. To upgrade the KVM, use the edgemicro upgradekvm command. For details on running this command, see Upgrading the KVM. You only need to do this step one time.
2. To upgrade the edgemicro-oauth proxy, use the edgemicro upgradeauth command. For details on running this command, see Ugrading the edgemicro-auth proxy. You only need to do this step one time.
3. Add the following line to your ~/.edgemicro/org-env-config.yaml file, where you must specify the same organization and environment that you configured the microgateway to use:

   jwk_public_keys: 'https://$ORG-$ENV.apigee.net/edgemicro-auth/jwkPublicKeys'

4. Run the key rotation command to rotate the keys. For details on this command, see Rotating keys.

   edgemicro rotatekey -o $ORG -e $ENV -k $KEY -s $SECRET

   For example:

   edgemicro rotatekey -o docs -e test \
   -k 27ee39567c75e4567a66236cbd4e86d1cc93df6481454301bd5fac4d3497fcbb \
   -s 4618b0008a6185d7327ebf53bee3c50282ccf45a3cceb1ed9828bfbcf1148b47
   

After key rotation, Edge returns multiple keys to Edge Microgateway. Note in the following example, each key has a unique "kid" (Key ID) value. The microgateway then uses these keys to validate authorization tokens. If the token validation fails, the microgateway looks to see if there is an older key in the key set and tries that key. The format of the returned keys is JSON Web Key (JWK). You can read about this format in RFC 7517.

{
  "keys": [
    {
      "kty": "RSA",
      "n": "nSl7R_0wKLiWi6cO3n8aOJwYGBtinq723Jgg8i7KKWTSTYoszOjgGsJf_MX4JEW1YCScwpE5o4o8ccQN09iHVTlIhk8CNiMZNPipClmRVjaL_8IWvMQp1iN66qy4ldWXzXnHfivUZZogCkBNqCz7VSC5rw2Jf57pdViULVvVDGwTgf46sYveW_6h8CAGaD0KLd3vZffxIkoJubh0yMy0mQP3aDOeIGf_akeZeZ6GzF7ltbKGd954iNTiKmdm8IKhz6Y3gLpC9iwQ-kex_j0CnO_daHl1coYxUSCIdv4ziWIeM3dmjQ5_2dEvUDIGG6_Az9hTpNgPE5J1tvrOHAmunQ",
      "e": "AQAB",
      "kid": "2"
    },
    {
      "kty": "RSA",
      "n": "8BKwzx34BMUcHwTuQtmp8LFRCMxbkKg_zsWD6eOMIUTAsORexTGJsTy7z-4aH0wJ3fT-3luAAUPLBQwGcuHo0P1JnbtPrpuYjaJKSZOeIMOnlryJCspmv-1xG4qAqQ9XaZ9C97oecuj7MMoNwuaZno5MvsY-oi5B_gqED3vIHUjaWCErd4reONyFSWn047dvpE6mwRhZbcOTkAHT8ZyKkHISzopkFg8CD-Mij12unxA3ldcTV7yaviXgxd3eFSD1_Z4L7ZRsDUukCJkJ-8qY2-GWjewzoxl-mAW9D1tLK6qAdc89yFem3JHRW6L1le3YK37-bs6b2a_AqJKsKm5bWw",
      "e": "AQAB",
      "kid": "1"
    }
  ]
}

Configuring a "not before" delay

For versions 3.1.5 and before, the new private key generated by the rotatekey command took effect immediately, and new tokens generated were signed with the new private key. However, the new public key was only made available to Edge Microgateway instances every 10 minutes (by default) when the microgateway configuration was refreshed. Because of this lag between the token signing and microgateway instance refresh, tokens signed with the latest key would be rejected until all instances received the public latest key.

In cases where multiple microgateway instances exist, the public key lag sometimes resulted in intermittent runtime errors with status 403, because token validation would pass on one instance, but fail on another instance until all instances were refreshed.

Starting in version 3.1.6, a new flag on the rotatekey command allows you to specify a delay for the new private key to become effective, allowing time for all microgateway instances to be refreshed and receive the new public key. The new flag is --nbf, which stands for "not before." This flag takes an integer value, the number of minutes to delay.

In the following example, the delay is set to 15 minutes:

edgemicro rotatekey -o docs -e test \
-k 27ee39567c75e4567a66236cbd4e86d1cc93df6481454301bd5fac4d3497fcbb \
-s 4618b0008a6185d7327ebf53bee3c50282ccf45a3cceb1ed9828bfbcf1148b47 \
--nbf 15

Note that a good practice is to set the delay to be more than the config_change_poll_internal configuration setting, which is 10 minutes by default. See also edgemicro attributes.

Filtering downloaded proxies

By default, Edge Microgateway downloads all of the proxies in your Edge organization that start with the naming prefix "edgemicro_". You can change this default to download proxies whose names match a pattern.

1. Open your Edge Micro config file: ~/.edgemicro/org-env-config.yaml
2. Add the proxyPattern element under edge_config. For example, the following pattern will download proxies such as edgemicro_foo, edgemicro_fast, and edgemicro_first.

   edge_config:
   …
   proxyPattern: edgemicro_f*

Specifying products without API proxies

In Apigee Edge, you can create an API product that does not contain any API proxies. This product configuration allows an API key associated with that product to work for with any proxy deployed in your organization. As of version 2.5.4, Edge Microgateway supports this product configuration.

Debugging and troubleshooting

Connecting to a debugger

You can run Edge Microgateway with a debugger, such as node-inspector. This is useful for troubleshooting and debugging custom plugins.

1. Restart Edge Microgateway in debug mode. To do this, add DEBUG=* to the beginning of the start command:

   DEBUG=* edgemicro start -o $ORG -e $ENV -k $KEY -s $SECRET

   To direct debug output to a file, you can use this command:

   export DEBUG=* nohup edgemicro start \
   -o $ORG -e $ENV -k $KEY -s $SECRET 2>&1 | tee /tmp/file.log

2. Start your debugger and set it to listen on the port number for the debugging process.
3. You can now step through the Edge Microgateway code, set breakpoints, watch expressions, and so on.

You can specify standard Node.js flags related to debug mode. For example, --nolazy helps with debugging asynchronous code.

Checking log files

If you're having problems, be sure to examine the log files for execution details and error information. For details, see Managing log files.

Using API key security

API keys provide a simple mechanism for authenticating clients making requests to Edge Microgateway. You can obtain an API key by copying the Consumer Key (also called Client ID) value from an Apigee Edge product that includes the Edge Microgateway authentication proxy.

Caching of keys

API keys are exchanged for bearer tokens, which are cached. You can disable caching by setting the Cache-Control: no-cache header on incoming requests to Edge Microgateway.

Using an API key

You can pass the API key in an API request either as a query parameter or in a header. By default, the header and query param name are both x-api-key.

Query parameter example:

curl http://localhost:8000/foobar?x-api-key=JG616Gjz7xs4t0dvpvVsGdI49G34xGsz

Header example:

curl http://localhost:8000/foobar -H "x-api-key:JG616Gjz7xs4t0dvpvVsGdI49G34xGsz"

Configuring the API key name

By default, x-api-key is the name used for both the API key header and query parameter. You can change this default in the configuration file, as explained in Making configuration changes. For example, to change the name to apiKey:

oauth:
  allowNoAuthorization: false
  allowInvalidAuthorization: false
  api-key-header: apiKey

In this example, both the query parameter and header name are changed to apiKey. The name x-api-key will no longer work in either case. See also Making configuration changes.

For example:

curl http://localhost:8000/foobar -H "apiKey:JG616Gjz7xs4t0dvpvVsGdI49G34xGsz"

For more information about using API keys with proxy requests, see Secure Edge Microgateway.

Enable upstream response codes

By default, the oauth plugin returns only 4xx error status codes if the response is not a 200 status. You can change this behavior so that it always returns the exact 4xx or 5xx code, depending on the error.

To enable this feature, add the oauth.useUpstreamResponse: true property to your Edge Microgateway configuration. For example:

oauth:
  allowNoAuthorization: false
  allowInvalidAuthorization: false
  gracePeriod: 10
  useUpstreamResponse: true

Using OAuth2 token security

This section explains how to get OAuth2 access tokens and refresh tokens. Access tokens are used to make secure API calls through the microgateway. Refresh tokens are used to obtain new access tokens.

How to get an access token

This section explains how to use the edgemicro-auth proxy to get an access token.

You can also get an access token using the edgemicro token CLI command. For details on the CLI, see Managing tokens.

API 1: Send credentials as body parameters

Substitute your org and environment names in the URL, and substitute the Consumer Id and Consumer Secret values obtained from a developer app on Apigee Edge for the client_id and client_secret body parameters:

curl -i -X POST "http://<org>-<test>.apigee.net/edgemicro-auth/token" \
-d '{"grant_type": "client_credentials", "client_id": "your_client_id", \
"client_secret": "your_client_secret"}' -H "Content-Type: application/json"

API 2: Send credentials in a Basic Auth header

Send the client credentials as a Basic Authentication header and the grant_type as a form parameter. This command form is also discussed in RFC 6749: The OAuth 2.0 Authorization Framework.

http://<org>-<test>.apigee.net/edgemicro-auth/token -v -u your_client_id:your_client_secret \
-d 'grant_type=client_credentials' -H "Content-Type: application/x-www-form-urlencoded"

Sample output

{
"token": "eyJraWQiOiIxIiwidHlwIjoi",
"access_token": "eyJraWQiOiIxIiwid",
"token_type": "bearer",
"expires_in": 1799
}

How to get a refresh token

To get a refresh token, make an API call to the /token endpoint of the edgemicro-auth proxy. You MUST make this API call with the password grant type. The following steps walk through the process.

1. Get an access and refresh token with the /token API. Note that the grant type is password:

   curl -X POST \
     https://your_organization-your_environment.apigee.net/edgemicro-auth/token \
     -H 'Content-Type: application/json' \
     -d '{
      "client_id":"mpK6l1Bx9oE5zLdifoDbF931TDnDtLq",
      "client_secret":"bUdDcFgv3nXffnU",
      "grant_type":"password",
      "username":"mpK6lBx9RoE5LiffoDbpF931TDnDtLq",
      "password":"bUdD2FvnMsXffnU"
   }'

   The API returns an access token and a refresh token. The response looks similar to this. Note that expires_in values integers and are specified in seconds.

   {
       "token": "your-access-token",
       "access_token": "your-access-token",
       "token_type": "bearer",
       "expires_in": 108,
       "refresh_token": "your-refresh-token",
       "refresh_token_expires_in": 431,
       "refresh_token_issued_at": "1562087304302",
       "refresh_token_status": "approved"
   }

2. You can now use the refresh token to get a new access token by calling the /refresh endpoint of the same API. For example:

   curl -X POST \
     https://willwitman-test.apigee.net/edgemicro-auth/refresh \
     -H 'Content-Type: application/json' \
     -d '{
      "client_id":"mpK6l1Bx9RoE5zLifoDbpF931TDnDtLq",
      "client_secret":"bUdDc2Fv3nMXffnU",
      "grant_type":"refresh_token",
      "refresh_token":"your-refresh-token"
   }'

   The API returns a new access token. The response looks similar to this:

   {
       "token": "your-new-access-token"
       }

Forever monitoring

Specifying a config file endpoint

If you run multiple Edge Microgateway instances, you may wish to manage their configurations from a single location. You can do this by specifying an HTTP endpoint where Edge Micro can download its configuration file. You can specify this endpoint when you start Edge Micro using the -u flag.

For example:

edgemicro start -o jdoe -e test -u http://mylocalserver/mgconfig -k public_key -s secret_key

where the mgconfig endpoint returns the contents of your configuration file. This is the file that, by default, is located in ~/.edgemicro and has the naming convention: org-env-config.yaml.

Disabling TCP connection data buffering

You can use the nodelay configuration attribute to disable data buffering for TCP connections used by Edge Microgateway.

By default TCP connections use the Nagle algorithm to buffer data before sending it off. Setting nodelay to true, disables this behavior (data will immediately fire off data each time socket.write() is called). See also the Node.js documentation for more details.

To enable nodelay, edit the Edge Micro config file as follows:

edgemicro:
  nodelay: true
  port: 8000
  max_connections: 1000
  config_change_poll_interval: 600
  logging:
    level: error
    dir: /var/tmp
    stats_log_interval: 60
    rotate_interval: 24

Running Edge Microgateway in standalone mode

You can run Edge Microgateway disconnected completely from any Apigee Edge dependency. This scenario, called standalone mode, lets you run and test Edge Microgateway without an Internet connection.

In standalone mode, the following features do not work, as they require connection to Apigee Edge:

• OAuth and API key
• Quota
• Analytics

On the other hand, custom plugins and spike arrest work normally, because they do not require a connection to Apigee Edge. In addition, a new plugin called extauth lets you authorize API calls to the microgateway with a JWT while in standalone mode.

Configuring and starting the gateway

To run Edge Microgateway in standalone mode:

1. Create a configuration file named as follows: $HOME/.edgemicro/$ORG-$ENV-config.yaml

   For example:

   vi $HOME/.edgemicro/foo-bar-config.yaml

2. Paste the following code into the file:

   edgemicro:
     port: 8000
     max_connections: 1000
     config_change_poll_interval: 600
     logging:
       level: error
       dir: /var/tmp
       stats_log_interval: 60
       rotate_interval: 24
     plugins:
       sequence:
         - extauth
         - spikearrest
   headers:
     x-forwarded-for: true
     x-forwarded-host: true
     x-request-id: true
     x-response-time: true
     via: true
   extauth:
     publickey_url: https://www.googleapis.com/oauth2/v1/certs
   spikearrest:
     timeUnit: second
     allow: 10
     buffersize: 0

3. Export the following environment variable with the value "1":

   export EDGEMICRO_LOCAL=1

4. Execute the following start command, where you provide values to instantiate the local proxy:

   edgemicro start -o $ORG -e $ENV -a $LOCAL_PROXY_NAME \
     -v $LOCAL_PROXY_VERSION -t $TARGET_URL -b $BASE_PATH

   Where:

   • $ORG is the "org" name that you used in the configuration file name.
   • $ENV is the "env" name that you used in the configuration file name.
   • $LOCAL_PROXY_NAME is the name of the local proxy that will be created. You can use any name you want.
   • $LOCAL_PROXY_VERSION is the version number for the proxy.
   • $TARGET_URL is the URL for the target of the proxy. (The target is the service that the proxy calls.)
   • $BASE_PATH is the base path of the proxy. This value must start with a forward slash. For a root base path, specify just a forward slash; for example, "/".

   For example:

   edgemicro start -o local -e test -a proxy1 -v 1 -t http://mocktarget.apigee.net -b /

5. Test the configuration.

   curl http://localhost:8000/echo  { "error" : "missing_authorization" }

   Because the extauth plugin is in the foo-bar-config.yaml file, you get a "missing_authorization" error. This plugin validates a JWT that must be present in the Authorization header of the API call. In the next section, you will obtain a JWT that will allow API calls to go through without the error.

Example: Obtaining an authorization token

The following example shows how to obtain a JWT from the Edge Microgateway JWT endpoint on Apigee Edge (edgemicro-auth/jwkPublicKeys). This endpoint is deployed when you perform a standard setup and configuration of Edge Microgateway. To obtain the JWT from the Apigee endpoint, you must first do the standard Edge Microgateway setup, and be connected to the Internet. The Apigee endpoint is used here for example purposes only and is not required. You can use another JWT token endpoint if you wish. If you do, then you'll need to obtain the JWT using the API provided for that endpoint.

The following steps explain how to get a token using the edgemicro-auth/jwkPublicKeys endpoint:.

1. You must perform a standard setup and configuration of Edge Microgateway to deploy the edgemicro-auth proxy to your organization/environment on Apigee Edge. If you did this step previously, you do not need to repeat it.
2. If you deployed Edge Microgateway to Apigee Cloud, you must be connected to the Internet so that you can obtain a JWT from this endpoint.
3. Stop Edge Microgateway:

   edgemicro stop

4. In the configuration file you created previously ($HOME/.edgemicro/org-env-config.yaml), point the extauth:publickey_url attribute to the edgemicro-auth/jwkPublicKeys endpoint in your Apigee Edge organization/environment. For example:

   extauth:
     publickey_url: 'https://your_org-your_env.apigee.net/edgemicro-auth/jwkPublicKeys'

5. Restart Edge Microgateway as you did previously, using the org/env names you used in the config file name. For example:

   edgemicro start -o foo -e bar -a proxy1 -v 1 -t http://mocktarget.apigee.net -b /

6. Get a JWT token from the authorization endpoint. Because you are using the edgemicro-auth/jwkPublicKeys endpoint, you can use this CLI command:

You can generate a JWT for Edge Microgateway using the edgemicro token command or an API. For example:

edgemicro token get -o your_org -e your_env \
  -i G0IAeU864EtBo99NvUbn6Z4CBwVcS2 -s uzHTbwNWvoSmOy

Where:

• your_org is the name of your Apigee organization for which you previously configured Edge Microgateway.
• your_env is an environment in the organization.
• The i option specifies the Consumer Key from a developer app that has a product that includes the edgemicro-auth proxy.
• The s option specifies the Consumer Secret from a developer app that has a product that includes the edgemicro-auth proxy.

This command asks Apigee Edge to generate a JWT that can then be used to verify API calls.

Test the standalone configuration

To test the configuration, call the API with the token added in the Authorization header as follows:

curl http://localhost:8000/echo -H "Authorization: Bearer your_token

Example:

curl http://localhost:8000/echo -H "Authorization: Bearer eyJraWQiOiIxIiwidHlwIjo...iryF3kwcDWNv7OQ"

Example output:

{
   "headers":{
      "user-agent":"curl/7.54.0",
      "accept":"*/*",
      "x-api-key":"DvUdLlFwG9AvGGpEgfnNGwtvaXIlUUvP",
      "client_received_start_timestamp":"1535134472699",
      "x-authorization-claims":"eyJhdDbiO...M1OTE5MTA1NDkifQ==",
      "target_sent_start_timestamp":"1535134472702",
      "x-request-id":"678e3080-a7ae-11e8-a70f-87ae30db3896.8cc81cb0-a7c9-11e8-a70f-87ae30db3896",
      "x-forwarded-proto":"http",
      "x-forwarded-host":"localhost:8000",
      "host":"mocktarget.apigee.net",
      "x-cloud-trace-context":"e2ac4fa0112c2d76237e5473714f1c85/1746478453618419513",
      "via":"1.1 localhost, 1.1 google",
      "x-forwarded-for":"::1, 216.98.205.223, 35.227.194.212",
      "connection":"Keep-Alive"
   },
   "method":"GET",
   "url":"/",
   "body":""
}

Using local proxy mode

In local proxy mode, Edge Microgateway does not require a microgateway-aware proxy to be deployed on Apigee Edge. Instead, you configure a "local proxy" by providing a local proxy name, basepath, and target URL when you start the microgateway. API calls to the microgateway are then sent to the target URL of the local proxy. In all other respects, local proxy mode works exactly the same as running Edge Microgateway in its normal mode. Authentication works the same, as do spike arrest and quota enforcement, custom plugins, and so on.

Use case and example

Local proxy mode is useful when you only need to associate one single proxy with an Edge Microgateway instance. For example, you can inject Edge Microgateway into Kubernetes as a sidecar proxy, where a microgateway and a service each run in a single pod, and where the microgateway manages traffic to and from its companion service. The following figure illustrates this architecture where Edge Microgateway functions as a sidecar proxy in a Kubernetes cluster. Each microgateway instance talks only to a single endpoint on its companion service:

A benefit of this style of architecture is that Edge Microgateway provides API management for individual services deployed to a container environment, such as a Kubernetes cluster.

Configuring local proxy mode

To configure Edge Microgateway to run in local proxy mode, follow these steps:

1. Run edgemicro init to set up your local configuration environment, exactly as you would in a typical Edge Microgateway setup. See also Configure Edge Microgateway.
2. Run edgemicro configure, as you would in a typical Edge Microgateway setup procedure. For example:

   edgemicro configure -o your_org -e your_env -u your_apigee_username

   This command deploys the edgemicro-auth policy to Edge and returns a key and secret that you will need to start the microgateway. If you need help, see Configure Edge Microgateway.

3. On Apigee Edge, create an API product and with the following mandatory configuration requirements (you can manage all other configurations as you wish):
   • You must add the edgemicro-auth proxy to the product. This proxy was deployed automatically when you ran edgemicro configure.
   • You must provide a resource path. Apigee recommends adding this path to the product: /**. To learn more, see Configuring the behavior of the resource path. See also Create API products in the Edge documentation.

4. On Apigee Edge, create a developer, or you can use an existing developer if you wish. For help, see Adding developers using the Edge management UI.

5. On Apigee Edge, create a developer app. You must add the API product you just created to the app. For help, see Registering an app in the Edge management UI.
6. On the machine where Edge Microgateway is installed, export the following environment variable with the value "1".

   export EDGEMICRO_LOCAL_PROXY=1

7. Execute the following start command:

   edgemicro start -o your_org -e your_environment -k your_key -s your_secret \
       -a local_proxy_name -v local_proxy_version -t target_url -b base_path

   Where:

   • your_org is your Apigee organization.
   • your_environment is an environment in your organization.
   • your_key is the key that was returned when you ran edgemicro configure.
   • your_secret is the secret that was returned when you ran edgemicro configure.
   • local_proxy_name is the name of the local proxy that will be created.
   • local_proxy_version is the version number for the proxy.
   • target_url is the URL for the target of the proxy (the service the proxy will call).
   • base_path is the base path of the proxy. This value must start with a forward slash. For a root base path, specify just a forward slash; for example, "/".

   For example:

   edgemicro start -o your_org -e test -k 7eb6aae644cbc09035a...d2eae46a6c095f \
     -s e16e7b1f5d5e24df...ec29d409a2df853163a -a proxy1 -v 1 \
     -t http://mocktarget.apigee.net -b /echo

Testing the configuration

You can test the local proxy configuration by calling the proxy endpoint. For example, if you specified a basepath of /echo, you can call the proxy as follows:

curl  http://localhost:8000/echo
{
  "error" : "missing_authorization",
  "error_description" : "Missing Authorization header"
}

This initial API call produced an error because you did not provide a valid API key. You can find the key in the Developer app you created previously. Open the app in the Edge UI, copy the Consumer Key, and use that key as follows:

curl  http://localhost:8000/echo -H 'x-api-key:your_api_key'

For example:

curl  http://localhost:8000/echo -H "x-api-key:DvUdLlFwG9AvGGpEgfnNGwtvaXIlUUvP"

Example output:

{
  "headers":{
    "user-agent":"curl/7.54.0",
    "accept":"*/*",
    "x-api-key":"DvUdLlFwG9AvGGpEgfnNGwtvaXIlUUvP",
    "client_received_start_timestamp":"1535134472699",
    "x-authorization-claims":"eyJhdWQiOi...TQ0YmUtOWNlOS05YzM1OTE5MTA1NDkifQ==",
    "target_sent_start_timestamp":"1535134472702",
    "x-request-id":"678e3080-a7ae-11e8-a70f-87ae30db3896.8cc81cb0-a7c9-11e8-a70f-87ae30db3896",
    "x-forwarded-proto":"http",
    "x-forwarded-host":"localhost:8000",
    "host":"mocktarget.apigee.net",
    "x-cloud-trace-context":"e2ac4fa0112c2d76237e5473714f1c85/1746478453618419513",
    "via":"1.1 localhost, 1.1 google",
    "x-forwarded-for":"::1, 216.98.205.223, 35.227.194.212",
    "connection":"Keep-Alive"
  },
  "method":"GET",
  "url":"/",
  "body":""
}

Using the synchronizer

This section explains how to use the synchronizer, an optional feature that improves the resiliency of Edge Microgteway by allowing it to retrieve configuration data from Apigee Edge and write it to a local Redis database. With a synchronizer instance running, other Edge Microgateway instances running on different nodes can retrieve their configuration directly from this database.

The syncrhonizer feature is currently supported to work with Redis 5.0.x.

What is the synchronizer?

The synchronizer provides a level of resilience for Edge Microgateway. It helps ensure that every instance of Edge Microgateway uses the same configuration, and that in the event of an internet disruption, Edge Microgateway instances can start up and run properly.

By default, Edge Microgateway instances must be able to communicate with Apigee Edge to retrieve and refresh their configuration data, such as API proxy and API product configurations. If the internet connection with Edge is disrupted, microgateway instances can continue to function because the latest configuration data is cached. However, new microgateway instances cannot start up without a clear connection. Furthermore, it is possible for an internet disruption to result in one or more microgateway instances running with configuration information that is out of sync with other instances.

The Edge Microgateway synchronizer provides an alternative mechanism for Edge Microgateway instances to retrieve configuration data that they require to start up and process API proxy traffic. The configuration data retrieved from calls to Apigee Edge include: the jwk_public_keys call, the jwt_public_key call, the bootstrap call, and the API products call. The synchronizer makes it possible for all of the Edge Microgateway instances running on different nodes to start up properly and stay in sync even if the internet connection between Edge Microgateway and Apigee Edge is disrupted.

The synchronizer is a specially configured instance of Edge Microgateway. Its only purpose is to poll Apigee Edge (the timing is configurable), retrieve configuration data, and write it to a local Redis database. The synchronizer instance itself cannot process API proxy traffic. Other instances of Edge Microgateway running on different nodes can be configured to retrieve configuration data from the Redis database rather than from Apigee Edge. Because all microgateway instances pull their configuration data from the local database, they can start up and process API requests even in the event of an internet disruption.

Configuring a synchronizer instance

Add the following configuration to the org-env/config.yaml file for the Edge Microgateway installation that you wish to use as the synchronizer:

edgemicro:
  redisHost: host_IP
  redisPort: host_port
  redisDb: database_index
  redisPassword: password
edge_config:
  synchronizerMode: 1
  redisBasedConfigCache: true

For example:

edgemicro:
  redisHost: 192.168.4.77
  redisPort: 6379
  redisDb: 0
  redisPassword: codemaster
edge_config:
  synchronizerMode: 1
  redisBasedConfigCache: true

Finally, save the configuration file and start the Edge Microgateway instance. It will begin polling Apigee Edge and storing downloaded configuration data in the Redis database.

Configuring regular Edge Microgateway instances

With the synchronizer running, you can configure additional Edge Microgateway nodes to run regular microgateway instances that process API proxy traffic. However, you configure these instances to obtain their configuration data from the Redis database rather than from Apigee Edge.

Add the following configuration to each additional Edge Microgateway node's org-env/config.yaml file. Note that the synchronizerMode property is set to 0. This property sets the instance to operate as a normal Edge Microgateway instance that processes API proxy traffic, and the instance will obtain its configuration data from the Redis database.

edgemicro:
  redisHost: host_IP
  redisPort: host_port
  redisDb: database_index
  redisPassword: password
edge_config:
  synchronizerMode: 0
  redisBasedConfigCache: true

For example:

edgemicro:
  redisHost: 192.168.4.77
  redisPort: 6379
  redisDb: 0
  redisPassword: codemaster
edge_config:
  synchronizerMode: 0
  redisBasedConfigCache: true

Configuration properties

The following configuration properties have been added to support the use of the synchronizer:

Configuring exclude URLs for plugins

You can configure the microgateway to skip the processing of plugins for specified URLs. You can configure these "exclude" URLs globally (for all plugins) or for specific plugins.

For example:

...
edgemicro:
  ...
  plugins:
    excludeUrls: '/hello,/proxy_one' # global exclude urls
    sequence:
      - oauth
      - json2xml
      - quota
json2xml:
  excludeUrls: '/hello/xml'  # plugin level exclude urls
...

In this example, plugins will not process incoming API proxy calls with the paths /hello or /proxy_one. In addition, the json2xml plugin will be skipped for APIs with /hello/xml in their path.

Setting configuration attributes with environment variable values

You can specify environment variables using tags in the configuration file. The specified environment variable tags are replaced by the actual environment variable values. Replacements are stored in memory only and not stored in the original configuration or cache files.

In this example, the attribute key is replaced by the value of the TARGETS_SSL_CLIENT_KEY environment variable, and so on.

targets:
  - ssl:
      client:
        key: <E>TARGETS_SSL_CLIENT_KEY</E>
        cert: <E>TARGETS_SSL_CLIENT_CERT</E>
        passphrase: <E>TARGETS_SSL_CLIENT_PASSPHRASE</E>

In this example, the <n> tag is used to indicate an integer value. Only positive integers are supported.

edgemicro:
  port: <E><n>EMG_PORT</n></E>

In this example, the <b> tag is used to indicate a boolean ( that is, true or false) value.

quotas:
  useRedis: <E><b>EMG_USE_REDIS</b></E>