Table of Contents
Add-on providers can improve the development experience of their Add-on by inserting data into the app’s log stream. This article explains how to connect your service to an app’s log stream via Logplex.
To gain access to the app’s Logplex token, which will give you the capability to write to an app’s log stream, you will first need to store the log drain token that is submitted in the Add-on provision request. More details of the provision request can be found in the Add-on Provider API Spec. If you have not previously stored the logplex_token from the provision request, you can use the App Info API to query for the token.
Data must be delivered to Logplex via HTTP. Using keep-alive connections and dense payloads, you can efficiently deliver all of your customer’s logs to Logplex via HTTP. Here is cURL example demonstrating a simple HTTP request:
$ curl "https://east.logplex.io/logs" \ --user "token:t123" \ -d "62 <190>1 2013-03-27T20:02:24+00:00 hostname t.123 procid - - foo62 <190>1 2013-03-27T20:02:24+00:00 hostname t.123 procid - - bar" \ -X "POST" \ -H "Content-Length: 130" \ -H "Content-Type: application/logplex-1"
Note that basic authentication is required. The username is
token and the password is the
logplex_token value returned by the App Info API.
Checkout all of the API details in the Logplex API Docs.
The headers of the request should contain Content-Length & Content-Type. The value of Content-Length should be the integer value of the length of the body. The value of Content-Type should be application/logplex-1.
The body of the HTTP request should contain length delimited syslog packets. Syslog packets are defined in RFC5424. The following line summarizes the RFC protocol:
<prival>version time hostname appname procid msgid structured-data msg
You can use
<190> (local7.info) for the prival and
1 for the version. The time field should be set to the time at which the logline was created in rfc3339 format. The hostname should be set to the name of your service. (e.g. postgresql) The appname field is reserved for the app’s Logplex token. The Logplex token is required to write to an app’s log stream. The procid is not used by logplex but is a great way to identify the process which is emitting the logs. Similarly, the msgid, and structured-data is not used by logplex and the value
- should be used. Finally, the msg is the section of the packet where the log message can be stored.
You should format your log messages in a way that is optimized for both human readability and machine parsability. With that in mind, log data should:
- Consist of a single message
- Use key-value pairs of the format
- Use a
sourcekey-value pair in log lines for distinguishing machines or environments (example:
- Show hierarchy with dots, from least to most specific, as in
- Units must immediately follow the number and must only include
These are log events that would benefit from statistical aggregation by a log consumer:
Periodically, your service may inject pre-aggregated metrics into a user’s logstream. The event could include the total number of database tables, active connections, cache usage:
sample#tables=30 sample#active-connections=3 sample#cache-usage=72.94 sample#cache-keys=1073002
Our experience logging for Postgres suggests that once a minute is a reasonable frequency for reporting aggregate metrics. Any higher frequency is potentially too noisy or expensive for storage/analysis. Be mindful of this when choosing to periodically log metrics from your add-on.
A reference implementation is provided at ryandotsmith/lpxc. This implementation highlights batching and keep-alive connections.