STATUS: this project is archived. You can replace usage of this drain with the syslog drain and send to an OpenTelemetry Collector using the syslog receiver.
Questions? You can chat with us in the Honeycomb Pollinators Slack. You can find a direct link to request an invite in Spread the Love: Appreciating Our Pollinators Community.
This is a log drain that can be added to any app running on Heroku to start observing that app's behaviour using Honeycomb.
It can send HTTP request events from the Heroku router, along with other events emitted from the Heroku platform. With some configuration it can also parse and send structured logs emitted by app code.
You need to set up your own instance of the log drain to receive logs from your app and send them to Honeycomb. The easiest way to do so is to deploy the drain to Heroku as well (as a separate app).
Clicking the following button will launch a Heroku UI that will deploy this drain as a new app in your Heroku account, prompting you along the way for configuration options.
Config is specified via environment variables (if the drain is deployed on Heroku, heroku config).
The drain requires the following config params:
HONEYCOMB_WRITE_KEY: the write key that identifies your Honeycomb team and enables you to send events. You can find this on your Honeycomb Account page.ALLOWED_APPS: authentication credentials for this drain, e.g.myapp:secretpassword. See Authentication.APP_FORMATS: specifies how to identify and parse the various types of logs emitted by your app, e.g.heroku/*:logfmt,app/*:raw. See Parsing logs.
The drain also supports the following optional config params:
HONEYCOMB_DATASET: set this to control which Honeycomb dataset this drain will send events to, e.g.myapporheroku.
- Once your instance of the drain is deployed, grab its url (e.g.
https://my-log-drain.herokuapp.com). - Add whatever credentials you configured in
ALLOWED_APPSto the URL as HTTP Basic Authentication so that your app can authenticate to the drain (e.g.https://myapp:[email protected]). heroku drains:add https://myapp:[email protected] --app=myapp
The drain uses HTTP Basic Authentication to ensure only your apps can send events to this log drain (and therefore to your Honeycomb dataset).
The authentication credentials are configured via the ALLOWED_APPS config param. The syntax of ALLOWED_APPS is a comma-delimited list of app:password pairs, e.g. myapp1:password1,myapp2:password2.
Heroku's logging infrastructure treats logs as unstructured strings, with a few bits of metadata to identify the source of each log entry (e.g. to distinguish logs emitted by app code from those emitted by the Heroku platform itself). Honeycomb works best with structured logs, so this drain supports configuring parsers for each log source to transform logs into structured data.
Log parsers are configured via the APP_FORMATS config param. The syntax of APP_FORMATS is a comma-delimited list of pattern:parser pairs, defined as follows.
Patterns are matched against the log source (proc_id field). Patterns can either be literal (e.g. heroku/router), or partial wildcards (e.g. app/* to match logs from any app process type). Wildcards are only supported after the slash (e.g. */* is not supported).
Parsers are identified by name, and must be one of the following:
logfmt: parses the logfmt "key1=value1 key2=value2" format (as used by the Heroku platform logs).json: parses JSON logs.raw: passes through unstructured text logs.ignore: ignores the matched logs.
If you are not sure what to set APP_FORMATS to, start with heroku/*:logfmt,app/*:raw. That will get you structured logs for things like HTTP requests, dyno CPU and RAM usage, database health metrics, and unstructured logs emitted from your app.
For the best experience, try configuring your app to emit logs as structured JSON, then replace app/*:raw with app/*:json in APP_FORMATS.