Logstream
Logstream is an Acquia service for streaming logs from Acquia Cloud. It supports streaming a variety of log sources:
- Web (Apache) access and error logs
- PHP error log
- Acquia Cloud's Drupal Request log
- Drupal watchdog log (if the syslog.module is enabled)
- Varnish cache logs
This repository contains a client-side library and CLI. Acquia Cloud provides a browser-based UI as well.
Quick start
- Logstream works in conjunction with Acquia's Cloud API. In order to use Cloud API you will need to generate credentials for your account. Cloud API v2 authentication explains how to generate the credentials. Place the credentials in
$HOME/.acquia/cloudapiv2.conf. This file is in JSON format with api_key and api_secret defined.
{
"api_key" : "key",
"api_secret" : "secret"
}
- Install the Logstream CLI:
$ gem install logstream
-
Find the UUID of the application you would like to stream logs for. Documentation on how to find applicaiton UUID can be found here
-
Stream logs from the production environment:
$ logstream tail 55ea1945-4aa6-4c56-bb7b-2108565e22d6 prod
127.0.0.1 - - [11/Jun/2014:17:28:47 +0000] "GET / HTTP/1.1" 200 7708 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.114 Safari/537.36" http_host=mysite.com affinity="-" upstream_addr="10.218.29.150:80" hosting_site=mysite request_time=0.030 forwarded_for="64.80.128.4" upstream_status="200"
... etc ...
A variety of filtering and display options are available:
$ logstream help tail
API v2
Logstream communicates over TCP using the WebSocket protocol. Use the logstream Cloud API call to retrieve the URL to connect to and an authenticated message to initial streaming for a particular Cloud environment.
Messages use text data frames, and contain JSON-encoded hashes. Each message contains a key 'cmd' which specifies the action to perform, and defines the other elements of the hash.
Available commands are defined in the following subsections. Inbound commands can arrive from the upstream client (end-user client, or an upstream logtailor). Outbound commands are sent to the upstream client.
In the examples that follow, "<-" indicates and inbound command and "->" indicates an outbound command. Those indicators are merely documentation for the examples; they are not part of the API.
API example
The logstream CLI --debug option shows API messages as they occur. Here is hypothetical debug output for the Dev environment of a site named "mysite" up to the first actual log message arriving. By default, the CLI enables the log types apache-request, php-error, drupal-watchdog, and varnish-request.
$ logstream tail 55ea1945-4aa6-4c56-bb7b-2108565e22d6 dev --debug
-> connect to wss://logstream.acquia.com/ah_websocket/logstream/v1
-> {"site":"devcloud:mysite","d":"deaefc1f42a4d18cb932c2eb9fa75115fba5ab83f1a3c564767ef1ce8dabf2cc","t":1404764927,"env":"dev","cmd":"stream-environment"}
<- {"cmd":"connected","server":"logstream-api-61"}
<- {"cmd":"connected","server":"bal-4"}
<- {"type":"bal-request","cmd":"available","server":"bal-4","display_type":"Balancer request"}
<- {"type":"varnish-request","cmd":"available","server":"bal-4","display_type":"Varnish request"}
-> {"cmd":"enable","type":"varnish-request","server":"bal-4"}
<- {"server":"srv-6","cmd":"connected"}
<- {"display_type":"Apache request","server":"srv-6","cmd":"available","type":"apache-request"}
-> {"cmd":"enable","type":"apache-request","server":"srv-6"}
<- {"display_type":"Apache error","server":"srv-6","cmd":"available","type":"apache-error"}
<- {"display_type":"PHP error","server":"srv-6","cmd":"available","type":"php-error"}
-> {"cmd":"enable","type":"php-error","server":"srv-6"}
<- {"display_type":"Drupal watchdog","server":"srv-6","cmd":"available","type":"drupal-watchdog"}
-> {"cmd":"enable","type":"drupal-watchdog","server":"srv-6"}
<- {"display_type":"Drupal request","server":"srv-6","cmd":"available","type":"drupal-request"}
<- {"server":"bal-5","cmd":"connected"}
<- {"msg":{"type":"varnish-request","server":"bal-4","cmd":"enable"},"cmd":"success","server":"bal-4"}
<- {"server":"srv-6","cmd":"success","msg":{"server":"srv-6","cmd":"enable","type":"apache-request"}}
<- {"server":"srv-6","cmd":"success","msg":{"server":"srv-6","cmd":"enable","type":"php-error"}}
<- {"server":"srv-6","cmd":"success","msg":{"server":"srv-6","cmd":"enable","type":"drupal-watchdog"}}
<- {"server":"bal-5","display_type":"Balancer request","cmd":"available","type":"bal-request"}
<- {"server":"bal-5","display_type":"Varnish request","cmd":"available","type":"varnish-request"}
-> {"cmd":"enable","type":"varnish-request","server":"bal-5"}
<- {"server":"bal-5","cmd":"success","msg":{"server":"bal-5","type":"varnish-request","cmd":"enable"}}
<- {"text":"107.0.255.129 - - [07/Jul/2014:20:28:53 +0000] \"GET / HTTP/1.0\" 200 2454 \"-\" \"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.153 Safari/537.36\" vhost=mysitedev.devcloud.acquia-sites.com host=mysitedev.devcloud.acquia-sites.com hosting_site=mysitedev pid=6863 request_time=80001 request_id=\"v-4fe9953a-0615-11e4-9fd8-1231392c7b9c\"","server":"srv-6","cmd":"line","http_status":200,"log_type":"apache-request","disp_time":"2014-07-07 20:28:53"}
107.0.255.129 - - [07/Jul/2014:20:28:53 +0000] "GET / HTTP/1.0" 200 2454 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.153 Safari/537.36" vhost=eebjaspandev.devcloud.acquia-sites.com host=eebjaspandev.devcloud.acquia-sites.com hosting_site=eebjaspandev pid=6863 request_time=80001 request_id="v-4fe9953a-0615-11e4-9fd8-1231392c7b9c"
success (outbound)
Sent by a variety of commands when a triggering command is completed successfully (failures result in an error command). Since some commands execute on multiple servers, a single command may generate multiple success replies.
Parameters:
- msg: the message that succeeded
Example:
<- {"cmd":"success","msg":{"type":"varnish-request","server":"bal-4","cmd":"enable"},"server":"bal-4"}
error (outbound)
Sent when an error occurs, either because of a bad request or any kind of system or network failure. Since some commands execute on multiple servers, a single command may generate multiple error replies.
Parameters:
- code: an HTTP-like status code; i.e.: 400 means your fault, 500 means our fault.
- description: a human-readable description of the error condition
- during: the operation that triggered the error
Example:
<- {"cmd"=>"error", "during"=>"{\"description\":\"this is an invalid message\"}", "description"=>"message missing cmd paramter", "code"=>400, "server"=>"logstream-api-61"}
connected (outbound)
Sent upon initial connection to a server.
Parameters:
- server: the name of the connected server
Example:
<- {"cmd":"connected","server":"logstream-api-61"}
available (outbound)
Sent to indicate an available log source.
Parameters:
- type: the log type (e.g. apache-access)
- display: the suggested display name for the log type (e.g. "Apache access")
- server: the originating server for this stream
Example:
<- {"type":"varnish-request","cmd":"available","server":"bal-4","display_type":"Varnish request"}
list-available (inbound)
Requests a list of all available log sources from one specified server or all servers. Parameters:
- server: the originating server to list; if not specified, all connected servers will reply.
One or all servers will send available messages for each available log source.
Example:
-> {"server":"srv-6","cmd":"list-available"}
<- {"display_type":"Apache request","server":"srv-6","cmd":"available","type":"apache-request"}
<- {"display_type":"Apache error","server":"srv-6","cmd":"available","type":"apache-error"}
enable (inbound)
Starts streaming a specific source previously offered via an "available" command.
Parameters:
- type: the log type to start streaming
- server: the originating server to stream from
Sends a success or error reply.
Example:
-> {"cmd":"enable","type":"varnish-request","server":"bal-4"}
disable (inbound)
Stops streaming from an enabled source.
Parameters:
- type: the log type to stop streaming
- server: the originating server to stream from
Sends a success or error reply.
Example:
-> {"cmd":"disable","type":"varnish-request","server":"bal-4"}
list-enabled (inbound) and enabled (outbound)
Requests a list of all enabled log sources from one specified server or all servers.
Parameters:
- server: the originating server to list; if not specified, all connected servers will reply.
One or all servers will send a enabled reply.
Parameters:
- enabled: an array of enabled log sources
- server: the server these logs are enabled on
Example:
-> {"server":"srv-6","cmd":"list-enabled"}
<-> {"server":"srv-6","cmd":"enabled","enabled":["drupal-watchdog","apache-request","php-error"]}
line (outbound)
Sent when a new, enabled line of log data is available. Parameters:
- type: the log type for this line (e.g. apache-access)
- server: this line's originating server name (e.g. web-1).
- unix_time: the Unix timestamp for this log line
- disp_time: the formatted display time for this log line
- text: the log text
Line messages may also contain other data depending on the log type:
- http_status: the HTTP status code, if the log line records an HTTP request (e.g. from Apache, nginx, Varnish, etc)
Example:
<- {"text":"107.0.255.129 - - [07/Jul/2014:20:28:53 +0000] \"GET / HTTP/1.0\" 200 2454 \"-\" \"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.153 Safari/537.36\" vhost=mysitedev.devcloud.acquia-sites.com host=mysitedev.devcloud.acquia-sites.com hosting_site=mysitedev pid=6863 request_time=80001 request_id=\"v-4fe9953a-0615-11e4-9fd8-1231392c7b9c\"","server":"srv-6","cmd":"line","http_status":200,"log_type":"apache-request","disp_time":"2014-07-07 20:28:53"}