v1.1.4

What's Changed

• Bump rexml from 3.2.6 to 3.2.8 by @dependabot in #24

Full Changelog: v1.1.3...v1.1.4

v1.1.3

What's Changed

• Cleanup gem setup by @stevenharman in #20

New Contributors

• @stevenharman made their first contribution in #20

Full Changelog: https://github.com/babbel/puma-plugin-telemetry/blob/main/CHANGELOG.md#113

v1.1.2

Make changes necessary for the gem to work with Puma v6.

1.1.1

First proper public release

1.1.0

Out of beta testing, reading for usage. Following is a recap from Alpha & Beta releases.

Added

• new metric: sockets.backlog (disabled by default), pulls information from Puma
  sockets about the state of their backlogs. This together with queue.backlog
  allows for full insights into total number of requests waiting to be processed
• config.sockets_telemetry! option to enable sockets telemetry
• config.socket_parser option to allow custom parser implementation as needed
• Datadog widgets examples under docs/examples.md

1.1.0.beta

Added

Different ways to parse Socket::Option. Mainly due to the fact that #inspect can't generate proper data on AWS Fargate, which runs Amazon Linux 2 with 4.14 kernel. So now besides #inspect there's also #unpack that parses binary data and picks proper field.

It depends on the kernel, but new fields are usually added at the end of the tcp_info struct, so it should more or less stay stable.

You can configure it by passing in config.socket_parser = :inspect or config.socket_parser = ->(opt) { your implementation }.

1.1.0.alpha

Added

Socket telemetry, and to be more precise new metric: sockets.backlog. If enabled it will
pull information from Puma sockets about the state of their backlogs (requests waiting to
be acknowledged by Puma). It will be exposed under sockets-backlog metric.

You can enable and test it via config.sockets_telemetry! option.

1.0.0

Added

• Release to Github Packages
• Explicitly flush datadog metrics after publishing them
• Middleware for measuring and tracking request queue time

Changed

• Replace statsd.batch with direct calls, as it aggregates metrics interally by default now.
  Also #batch method is deprecated and will be removed in version 6 of Datadog Statsd client.

0.3.1

Changed

• IO target replaces dots in telemetry keys with dashes for better integration with AWS CloudWatch

0.3.0

Added

• Datadog Target integration tests

Fixed

• Datadog Target