Gravity Log Application

Gravity Log Application is intended for gathering logs and providing API for logs retrieval in Gravity cluster. The application is based on Logrange streaming database.

1. HTTP API

The server listens on port 8083 by default.

Querying Logs

GET: /v1/log?limit=&query=

• query can contain the following terms:
  

  • pod: - to limit search to a specific pod
    
  • container: - to limit search to a specific container inside a pod
    
  • file: - to limit search to a specific log file
    
  • boolean operators and and or

  Example:
  /v1/log?limit=1&query=pod:p1 and container:"c1" and file:f1 or file:f2

• query param could be used to search for literal text occurrence:

  Example:
  /v1/log?limit=100&query="some text"

• default limit is 1000

• output format: JSON, e.g.: [{"type":"data","payload":""}]

Download logs

GET: /v1/download

• output format: compressed tarball stream (tar.gz)

2. Recurring jobs

The application has a set of recurring jobs which run as a part of the application binary and perform tasks described below.

Config synchronizer

Job that syncs the updates to Gravity log forwarder configuration to internal format. Frequency can be configured (see SyncIntervalSec param of the config) and is set to 20 seconds by default.

Queries executor

Job that runs scheduled queries. Queries can be configured (see CronQueries section of the config), by default there is a single query configured which is used to keep the database size within limits by periodically trimming older entries.

3. Build instructions

 $ git clone git@github.com:gravitational/logging-app.git
 $ cd logging-app
 $ make clean tarball

Prefix the make command with OPS_URL=ops-url to work with a remote package repository and VERSION=x.y.z to build a specific version.