-
Notifications
You must be signed in to change notification settings - Fork 530
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Tweak telemetry events on batch middleware #1170
base: main
Are you sure you want to change the base?
Tweak telemetry events on batch middleware #1170
Conversation
Add instrumentation for `Absinthe`, the GraphQL Elixir-framwork. Relies on some new telemetry work on Absinthe: * [Added telemetry for async helper][1], used to propagate traces * [Improved telemetry for batch helper][2], used to propagate traces * [Tweaked pipeline instrumentation][3], that splits parse/validate/execute phases, following the approach of [JavaScript libraries][4] Span names and attributes also follow the [JavaScript impl][4], the closes to a future [Semantic Attribute][5] spec. The "span tracker" used here is to connect traces in a shape that resembles the query shape instead of execution shape. One example is: ``` query { users { name } } ``` In case `name` is an async resolver (for example), the execution will resembles a recursive function, and such our trace will be: ``` RootQueryType.users User.name User.name User.name ``` With the span tracker, the above trace will become: ``` RootQueryType.users User.name User.name User.name ``` That helps visually scan the traces, and it's friendly to folks who doesn't know much about Absinthe internals. [1]: absinthe-graphql/absinthe#1169 [2]: absinthe-graphql/absinthe#1170 [3]: absinthe-graphql/absinthe#1172 [4]: https://www.npmjs.com/package/@opentelemetry/instrumentation-graphql [5]: open-telemetry/opentelemetry-specification#2456
With a similar rationale from [this PR][1], tweak the telemetry events from batch middleware. The existing event contract stays the same, but move to run from inside the Task used for batch execution. There is a small semantic difference of when the :stop event fires - this patch sends the event at the end of the task, while the current form may have a delay/sync caused by Task await. A complementary event is added as well to announce when the Resolver continuation - the post batch - happens. Given that piece is part of the resolver, the resolution struct is included as part of the metadata. [1]: absinthe-graphql#1169
f873819
to
719e54c
Compare
Add docs for both Async and Batch events, covering the recent changes made in relation to both sets. The docs are written following conventions found in many libraries that adopted `:telemetry` since the first implemention done at Absinthe. The original text wasn't touched, but there is room for improvement there as well. Given the amount of events, some grouping was done by "middleware".
719e54c
to
8c36f75
Compare
:telemetry.span([:absinthe, :middleware, :async, :task], %{}, fn -> {fun.(), %{}} end) | ||
:telemetry.span([:absinthe, :middleware, :async, :task], %{}, fn -> | ||
result = fun.() | ||
{result, %{result: result}} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry for this non-related change, but as I've wrote the telemetry docs I felt like could be a good consistency change as batch telemetry events returns results as well.
Add docs for both Async and Batch events, covering the recent changes made in relation to both sets. The docs are written following conventions found in many libraries that adopted Given the amount of events, I've grouped them by middleware, and phrased sections so Any thoughts @benwilson512? Thanks in advance! |
"erlex": {:hex, :erlex, "0.2.6", "c7987d15e899c7a2f34f5420d2a2ea0d659682c06ac607572df55a43753aa12e", [:mix], [], "hexpm", "2ed2e25711feb44d52b17d2780eabf998452f6efda104877a3881c2f8c0c0c75"}, | ||
"ex_doc": {:hex, :ex_doc, "0.27.3", "d09ed7ab590b71123959d9017f6715b54a448d76b43cf909eb0b2e5a78a977b2", [:mix], [{:earmark_parser, "~> 1.4.19", [hex: :earmark_parser, repo: "hexpm", optional: false]}, {:makeup_elixir, "~> 0.14", [hex: :makeup_elixir, repo: "hexpm", optional: false]}, {:makeup_erlang, "~> 0.1", [hex: :makeup_erlang, repo: "hexpm", optional: false]}], "hexpm", "ee60b329d08195039bfeb25231a208749be4f2274eae42ce38f9be0538a2f2e6"}, | ||
"ex_doc": {:hex, :ex_doc, "0.28.4", "001a0ea6beac2f810f1abc3dbf4b123e9593eaa5f00dd13ded024eae7c523298", [:mix], [{:earmark_parser, "~> 1.4.19", [hex: :earmark_parser, repo: "hexpm", optional: false]}, {:makeup_elixir, "~> 0.14", [hex: :makeup_elixir, repo: "hexpm", optional: false]}, {:makeup_erlang, "~> 0.1", [hex: :makeup_erlang, repo: "hexpm", optional: false]}], "hexpm", "bf85d003dd34911d89c8ddb8bda1a958af3471a274a4c2150a9c01c78ac3f8ed"}, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Took the opportunity to update the library. If there are other preferred means to do so, I'm glad to revert.
The `:monotonic_time` is part of [1.1.0 release of :telemetry][1]. Enforce that version at minimum to match documented Telemetry for Async and Batch middleware events. [1]: https://github.com/beam-telemetry/telemetry/blob/091121f6153840fd079e68940715a5c35c5aa445/CHANGELOG.md#110
8de92a3
to
05c7318
Compare
With a similar rationale from this PR, tweak the telemetry events from batch middleware.
The existing event contract stays the same, but move to run from inside the Task used for batch execution. There is a small semantic difference of when the :stop event fires - this patch sends the event at the end of the task, while the current form may have a delay/sync caused by Task await.
A complementary event is added as well to announce when the Resolver continuation - the post batch - happens. Given that piece is part of the resolver, the resolution struct is included as part of the metadata.