statsrelay (3.0)

A veritable toolkit, sidecar, or daemon(set) for sharding, aggregating, relaying and working with statsd and Prometheus based metrics sources, at scale.

License

MIT License Copyright (c) 2015-2020 Lyft Inc.

Originally based on statsrelay: Copyright (c) 2014 Uber Technologies, Inc.

Whats different in this version

Statsrelay 3.0 is a port of the original C statsrelay to Rust, with a number of new features designed to improve the operatability and scalability of the original daemon, moving beyond pure "relaying" and instead focusing on both sharding, as well as cascading aggregation. The original C daemon in this fork featured sampling support, but was limited by having the output format be statsd.

Build

Dependencies:

• Rust (stable, 1.44+)

Use

statsrelay 3.0.0

USAGE:
    statsrelay [OPTIONS]

FLAGS:
    -h, --help       Prints help information
    -V, --version    Prints version information

OPTIONS:
    -c, --config <config>     [default: /etc/statsrelay.json]

Statsrelay logging is handled by the env_logger crate, which inherits a number of logging options from the environment. Consult the crate documentation for more information on options you can set.

Protocols

Statsrelay understands:

• statsd text line protocol
  • with sampling support (@sampling)
  • with extended data types (map, kv, sets, etc)
  • with "DogStatsD" extended tags (|#tags)
  • with Lyft internal tags (metric.__tag=value)

Configuration file

The configuration file is a JSON file originating from the original statsrelay structure. Currently, any valid Lyft statsrelay configuration file is accepted as a statsrelay file.

Basic structure

{
    "statsd": {
        "bind": "127.0.0.1:8129",
        "shard_map": [
            "127.0.0.1:8122"
        ],
        "validate": true,
        "prefix": "myapp.",
        "suffix": ".suffix",
        duplicate_to: []
    }
}

Statsd inputs and routing is defined in the outer statsd block.

• bind: sets the server bind address to accept statsd protocol messages. Statsrelay will bind on both UDP and TCP ports.
• shard_map: defines where to send statsd output to, from a list of servers. The same server can be specified more than once (allowing for virtual sharding). Output statsd lines are consistently hashed, and sent to the corresponding server based on a standard hash ring, in a compatible format to the original statsrelay code (Murmur3 hash). This list can be empty to not relay statsd messages.
• validate: turns on extended, more expensive validation of statsd line protocol messages, such as parsing of numerical fields, which may not be required for a pure relaying case.
• prefix: prepend this prefix string in front of every metric/statsd line before forwarding it to the shard_map servers. Useful for tagging metrics coming from a sidecar.
• suffix: append a suffix. Works like prefix, just at the end.
• duplicate_to forks the incoming statsd metrics down a number of parallel processing pipelines.

duplicate_to options and sampling