[DI] Batch outgoing http requests

[watson]

What does this PR do?

Adds a queue in front of the outgoing http requests from the Dynamic Instrumentation devtools worker thread.

Two queues are added:

• One for requests to /debugger/v1/diagnostics
• One for requests to /debugger/v1/input

Both queues are configured in the same manner.

When adding a payload to any of the queues, it's buffered until either a 1 second timer ends or the size of the queue exceeds 5MB after which the queue is flushed. Once flushed, all payloads waiting in the queue are sent to the HTTP endpoint in a single HTTP request.

Motivation

Without this queue the tracer could produce up to 5.000 request per second. This would overload the tracer.

Note to reviewers

Hide whitespace changes for a better diff

[watson]

• [DI] Add Sirun benchmark for Dynamic Instrumentation #5004
• [DI] Batch outgoing http requests #5007 👈 (View in Graphite)
• master

This stack of pull requests is managed by Graphite. Learn more about stacking.

[github-actions]

Overall package size

Self size: 8.42 MB
Deduped: 94.88 MB
No deduping: 95.4 MB

Dependency sizes

| name | version | self size | total size | |------|---------|-----------|------------| | @datadog/libdatadog | 0.3.0 | 29.43 MB | 29.43 MB | | @datadog/native-appsec | 8.3.0 | 19.37 MB | 19.38 MB | | @datadog/native-iast-taint-tracking | 3.2.0 | 13.9 MB | 13.91 MB | | @datadog/pprof | 5.4.1 | 9.76 MB | 10.13 MB | | protobufjs | 7.2.5 | 2.77 MB | 5.16 MB | | @datadog/native-iast-rewriter | 2.6.1 | 2.59 MB | 2.73 MB | | @opentelemetry/core | 1.14.0 | 872.87 kB | 1.47 MB | | @datadog/native-metrics | 3.1.0 | 1.06 MB | 1.46 MB | | @opentelemetry/api | 1.8.0 | 1.21 MB | 1.21 MB | | import-in-the-middle | 1.11.2 | 112.74 kB | 826.22 kB | | source-map | 0.7.4 | 226 kB | 226 kB | | opentracing | 0.14.7 | 194.81 kB | 194.81 kB | | lru-cache | 7.18.3 | 133.92 kB | 133.92 kB | | pprof-format | 2.1.0 | 111.69 kB | 111.69 kB | | @datadog/sketches-js | 2.1.0 | 109.9 kB | 109.9 kB | | semver | 7.6.3 | 95.82 kB | 95.82 kB | | lodash.sortby | 4.7.0 | 75.76 kB | 75.76 kB | | ignore | 5.3.1 | 51.46 kB | 51.46 kB | | shell-quote | 1.8.1 | 44.96 kB | 44.96 kB | | istanbul-lib-coverage | 3.2.0 | 29.34 kB | 29.34 kB | | rfdc | 1.3.1 | 25.21 kB | 25.21 kB | | @isaacs/ttlcache | 1.4.1 | 25.2 kB | 25.2 kB | | tlhunter-sorted-set | 0.1.0 | 24.94 kB | 24.94 kB | | limiter | 1.1.5 | 23.17 kB | 23.17 kB | | dc-polyfill | 0.1.4 | 23.1 kB | 23.1 kB | | retry | 0.13.1 | 18.85 kB | 18.85 kB | | jest-docblock | 29.7.0 | 8.99 kB | 12.76 kB | | crypto-randomuuid | 1.0.0 | 11.18 kB | 11.18 kB | | path-to-regexp | 0.1.12 | 6.6 kB | 6.6 kB | | koalas | 1.0.2 | 6.47 kB | 6.47 kB | | module-details-from-path | 1.0.3 | 4.47 kB | 4.47 kB |

_{🤖 This report was automatically generated by heaviest-objects-in-the-universe}

[pr-commenter]

Benchmarks

Benchmark execution time: 2024-12-20 18:19:06

Comparing candidate commit 39ba330 in PR branch watson/DEBUG-3228/batch-requests with baseline commit 3798033 in branch master.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 782 metrics, 16 unstable metrics.

[watson]

The system tests are now failing because we're sending an array instead of an object to the endpoints. Previously we were ignoring anything that wasn't an array, but now we start to validate the actual content, which means we're now seeing errors we wasn't seeing before.

We see two types of errors, both indicating a missing required field in the diagnostics messages: probeVersion and message. I belive these are actually because of some issues with the system-tests themselves and not this repo, so I've created a couple of PRs to fix the errors there:

• [Debugger] Ensure all probes have a version property system-tests#3581
• [Debugger] Remove message field from diagnostics schema system-tests#3667

[shatzi]

This is great. thank you.

currently we only creating snapshots probes. but once we have log probes we should have another take on this to handle chatty probes. Would be great if we can unify the spec for that before we implement.

[shatzi]

I think it better to pass just the json and remove the size param.
If the user give wrong values it break the logic of the queue

[watson]

This is purely a performance optimization. In the snapshot code, the size is already calculated since we need to know if it's larger than 1MB. So to not calculate it twice we allow it to be passed in. It's not user generated input.