Repository navigation
test: stabilize flaky tests - #2256
Conversation
BenchmarksComparisonBenchmark execution time: 2026-07-21 10:41:53 Comparing candidate commit 059df57 in PR branch Found 0 performance improvements and 0 performance regressions! Performance is the same for 123 metrics, 0 unstable metrics.
|
🎉 All green!🧪 All tests passed 🎯 Code Coverage (details) 🔗 Commit SHA: 059df57 | Docs | Datadog PR Page | Give us feedback! |
yannham
left a comment
There was a problem hiding this comment.
If this unblocks, let's merge this 👍
| } | ||
|
|
||
| #[test] | ||
| #[cfg_attr(miri, ignore)] |
There was a problem hiding this comment.
I'm curious, why would miri behave differently wrt to flakiness? Isn't just a red herring (limiting the probability of hitting the issue if we run the test twice as less? Is it a new test that just fails under Miri consistently?
There was a problem hiding this comment.
So I could not repro this. Clanker has theories about the test that could be an issue:
For this specific test:
enumerate_discovers_spawned_threadthe test spawns a thread, receives its TID over a channel, then enumerates /proc/self/task and expects that TID to be present.
The channel guarantees the child thread sent its TID, but it does not necessarily prove that Miri/host /proc view will expose that interpreted thread exactly like a normal native Linux thread at that point. Under Miri, threads are simulated/interpreted, while /proc
is an OS interface. With -Zmiri-disable-isolation, Miri allows more host access, but that does not make every host-observable thread assumption robust.
Artifact Size Benchmark Reportaarch64-alpine-linux-musl
aarch64-unknown-linux-gnu
libdatadog-x64-windows
libdatadog-x86-windows
x86_64-alpine-linux-musl
x86_64-unknown-linux-gnu
|
| with: | ||
| cache-targets: true # cache build artifacts | ||
| cache-bin: true # cache the ~/.cargo/bin directory | ||
| - name: "Clear macOS panic-abort cache" |
There was a problem hiding this comment.
@gyuheon0h this is a workaround, I think we should not be caching this.
|
/merge |
|
View all feedbacks in Devflow UI.
This pull request is not mergeable according to GitHub. Common reasons include pending required checks, missing approvals, or merge conflicts — but it could also be blocked by other repository rules or settings.
The expected merge time in
|
054402d
into
main
…er-flare (#2332) # Release proposal for libdd-ffe, libdd-http-client, libdd-tracer-flare and their dependencies This PR contains version bumps based on public API changes and commits since last release. ## libdd-capabilities **Next version:** `3.0.0` **Semver bump:** `major` **Tag:** `libdd-capabilities-v3.0.0` ### Commits - fix(telemetry)!: stop sending Connection: close to the Agent (#2286) - feat(capabilities)!: Add streaming to http capabilities (#2251) - feat: added file capability [APMSP-3780] (#2240) - feat: added environment capability [APMSP-3780] (#2239) ## libdd-common **Next version:** `5.1.1` **Semver bump:** `patch` **Tag:** `libdd-common-v5.1.1` ### Commits - feat(telemetry)!: Add Installation signature and AppProduct changes payloads (#2213) - refactor: migrate to workspace dependencies, phase 4 (#2296) - refactor: migrate to workspace dependencies, phase 3 (#2283) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) - test: stabilize flaky tests (#2256) - fix(sidecar): don't double-encode file:// telemetry endpoints (#2230) - fix: last new clippy lint (#2221) ## libdd-tinybytes **Next version:** `1.1.2` **Semver bump:** `patch` **Tag:** `libdd-tinybytes-v1.1.2` ### Commits - chore: moving to workspace-level dependencies, phase 2 (#2270) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) ## libdd-trace-protobuf **Next version:** `4.0.1` **Semver bump:** `patch` **Tag:** `libdd-trace-protobuf-v4.0.1` ### Commits - refactor: migrate to workspace dependencies, phase 4 (#2296) - chore: moving to workspace-level dependencies, phase 2 (#2270) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) - feat(stats)!: implement additional metric tags in libdd-trace-stats (#2170) ## libdd-capabilities-impl **Next version:** `4.0.0` **Semver bump:** `major` **Tag:** `libdd-capabilities-impl-v4.0.0` ###⚠️ major bump forced due to: - `libdd-capabilities`: ^2.1.0 → ^3.0.0 ### Commits - feat(telemetry)!: Add Installation signature and AppProduct changes payloads (#2213) - refactor: migrate to workspace dependencies, phase 4 (#2296) - fix(telemetry)!: stop sending Connection: close to the Agent (#2286) - fix(remote-config)!: Expose HttpClientCapability in remote config (#2252) - feat(capabilities)!: Add streaming to http capabilities (#2251) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) - feat: added file capability [APMSP-3780] (#2240) - feat(telemetry)!: make telemetry worker wasm-compatible for the TraceExporter (#2172) - feat: added environment capability [APMSP-3780] (#2239) ## libdd-http-client **Next version:** `1.0.0` **Semver bump:** `major` **Tag:** `libdd-http-client-v1.0.0` **Warning:** this is an initial release. Please verify that the version and commits included are correct. ## libdd-trace-normalization **Next version:** `3.0.1` **Semver bump:** `patch` **Tag:** `libdd-trace-normalization-v3.0.1` ### Commits - refactor: migrate to workspace dependencies, phase 4 (#2296) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) ## libdd-remote-config **Next version:** `3.0.0` **Semver bump:** `major` **Tag:** `libdd-remote-config-v3.0.0` ### Commits - refactor(rc)!: make conversion from RemoteConfigProduct back and forth generally available (#2325) - fix(wasm)!: finish the WASM port of remote-config (#2315) - refactor: migrate to workspace dependencies, phase 3 (#2283) - feat(rc): Handle expired config status (#2274) - feat(remote-config)!: add AsmRawResponseBody capability (#2278) - chore: moving to workspace-level dependencies, phase 2 (#2270) - fix(remote-config)!: Expose HttpClientCapability in remote config (#2252) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) - fix(rc): Make Target fields available again after eaf5ad0 (#2232) - fix: new clippy lints (#2219) ## libdd-trace-utils **Next version:** `10.0.0` **Semver bump:** `major` **Tag:** `libdd-trace-utils-v10.0.0` ###⚠️ major bump forced due to: - `libdd-capabilities`: ^2.1.0 → ^3.0.0 ### Commits - refactor(rc)!: make conversion from RemoteConfigProduct back and forth generally available (#2325) - fix(libdd-trace-utils): apply SpanLink flags masking when v0.5 json encoding (#2314) - refactor: migrate to workspace dependencies, phase 4 (#2296) - feat(data-pipeline)!: add compression option for agentless export (#2205) - refactor: migrate to workspace dependencies, phase 3 (#2283) - refactor(trace-utils)!: split generic TracerHeaderTags (#2279) - chore(lint): update nightly (#2281) - chore: moving to workspace-level dependencies, phase 2 (#2270) - fix(remote-config)!: Expose HttpClientCapability in remote config (#2252) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) - feat(stats)!: implement additional metric tags in libdd-trace-stats (#2170) - feat(trace-utils)!: add v1 decoder (#2174) - feat(trace-utils)!: add encoder v1 to v04 + refactor (#2145) - feat(otlp): add instrumentation scope metadata (#2235) - fix(trace-utils)!: serialize v0.5 span links and events into meta (#980) ## libdd-ffe **Next version:** `1.0.0` **Semver bump:** `major` **Tag:** `libdd-ffe-v1.0.0` **Warning:** this is an initial release. Please verify that the version and commits included are correct. ## libdd-tracer-flare **Next version:** `1.0.0` **Semver bump:** `major` **Tag:** `libdd-tracer-flare-v1.0.0` **Warning:** this is an initial release. Please verify that the version and commits included are correct. [APMSP-3780]: https://datadoghq.atlassian.net/browse/APMSP-3780?atlOrigin=eyJpIjoiNWRkNTljNzYxNjVmNDY3MDlhMDU5Y2ZhYzA5YTRkZjUiLCJwIjoiZ2l0aHViLWNvbS1KU1cifQ [APMSP-3780]: https://datadoghq.atlassian.net/browse/APMSP-3780?atlOrigin=eyJpIjoiNWRkNTljNzYxNjVmNDY3MDlhMDU5Y2ZhYzA5YTRkZjUiLCJwIjoiZ2l0aHViLWNvbS1KU1cifQ [APMSP-3780]: https://datadoghq.atlassian.net/browse/APMSP-3780?atlOrigin=eyJpIjoiNWRkNTljNzYxNjVmNDY3MDlhMDU5Y2ZhYzA5YTRkZjUiLCJwIjoiZ2l0aHViLWNvbS1KU1cifQ --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: hoolioh <107922352+hoolioh@users.noreply.github.com>
# Release proposal for libdd-crashtracker, libdd-gotter and their dependencies This PR contains version bumps based on public API changes and commits since last release. ## libdd-gotter **Next version:** `1.0.0` **Semver bump:** `major` **Tag:** `libdd-gotter-v1.0.0` **Warning:** this is an initial release. Please verify that the version and commits included are correct. ## libdd-crashtracker **Next version:** `2.0.0` **Semver bump:** `major` **Tag:** `libdd-crashtracker-v2.0.0` ###⚠️ major bump forced due to: - `libdd-common`: ^1.0.0 → ^5.2.0 - `libdd-telemetry`: ^1.0.0 → ^7.0.0 ### Commits - feat(crashtracking)!: retrieve c assert message for linux when `__assert_fail` is dynamically loaded (#2268) - test: skip/shorten slow miri jobs (#2331) - feat(crashtracking): send debug log when no data is received at all (#2321) - fix(crashtracking): check fields and exclude uuid for `has_data` (#2322) - refactor: migrate to workspace dependencies, phase 4 (#2296) - refactor: migrate to workspace dependencies, phase 3 (#2283) - chore: moving to workspace-level dependencies, phase 2 (#2270) - refactor: consolidate core dependencies at workspace level (phase 1) (#2253) - test: stabilize flaky tests (#2256) - feat(telemetry)!: make telemetry worker wasm-compatible for the TraceExporter (#2172) - chore(crashtracking): harden multi thread ptrace collection (#2216) - fix(sidecar): don't double-encode file:// telemetry endpoints (#2230) - fix: new clippy lints (#2219) - refactor(sidecar)!: Avoid a dedicated socket for crashtracker (#2179) - refactor(clippy): prefer core and alloc imports (#2196) - refactor(libdd-telemetry)!: avoid leaking libdd-common types in the public API (#2152) - feat(crashtracking): collect all native stacks for unhandled exception and also crashing thread (#2155) - fix(crashtracking): sanitize type and message for unhandled exceptions (#2148) - fix(crashtracking): multi thread collection centos flakes harden (#2113) - chore(crashtracking)!: remove frame count experimental field (#2114) - fix(crashtracking): authenticate peer granted socket ptrace access (#2098) - feat(crashtracking)!: add experimental frame count field (#2099) - fix(crashtracker): increase test_waitall_nohang timeout to 500ms (#2097) - fix(crashtracker): support socket based receiver for all thread collection (#2080) - fix(crashtracker): set failed thread stack collection as incomplete empty stack (#2079) - fix(crashtracking)!: flatten all threads object into a list of `ThreadData` (#2054) - refactor(crashtracker): update imports to linux only (#2036) - fix(crashtracker): move preload logger marking after recursive guard (#2023) - fix(libdd-telemetry): restore previous Cargo.toml version (#1993) - feat!: integrate obfuscation to the stats exporter [APMSP-2764] (#1819) - chore(crashtracking)!: create errorsintake crash ping directly from telemetry (#1963) - feat(telemetry)!: include dependencies and integrations in app-extended-heartbeat (#1962) - feat(crashtracking)!: collect all threads (#1878) - chore(crashtracking): bump libdatadog-libunwind to v1.0.2 (#1942) - chore(crashtracking): default errors intake crash report upload to be on (#1902) - chore(libdd-crashtracker): remove path reference for libdd-libunwind-sys (#1877) - fix(crashtracking): handle new lines in client submitted exception message (#1836) - fix(crashtracker): Fix bin_tests in gitlab (#1832) - feat(crashtracking)!: improve parity between errors intake payload and telemetry intake payload (#1823) - feat(crashtracking)!: emit ucontext registers as structured data (#1787) - fix(ci): use single threaded for all tests that mutate signal state and use non-fatal signal (#1812) - fix(crashtracking): fix SIGCHLD signal guarding while in CT signal handler (#1807) - fix(ci): use single threaded to avoid race conditions for sa guard tests (#1800) - fix(crashtracking): guard sigchld and sigpipe during crashtracker signal handler execution (#1771) - chore(crashtracking): preserve errno for crashtracker (#1767) - chore(crashtracker): use weaker mem ordering for OP_COUNTERS (#1744) - chore(crashtracking): rename target triple to runtime platform (#1747) - chore(ci): give libdd-libunwind-sys its own version (#1743) - chore(crashtracking): add tag for target triple (#1741) - refactor(libdd-crashtracker)!: avoid leaking Endpoint through the public API (#1705) - chore(cargo): bump to 29.0.0 (#1702) - fix(crashtracking): use libunwind to unwind frames (#1663) - chore(deps): bump blazesym to 0.2.3 and blazesym-c to 0.1.7 (#1654) - chore(ci): fix crashtracker receiver binary rpath setting (#1652) - chore(crashtracking): emit a best effort stacktrace for Mac (#1645) - chore(crashtracker): use default-features=false for aws-lc-sys (#1625) - feat(crashtracking): report unhandled exceptions (#1596) - refactor(ddcommon)!: remove direct dependency on hyper client everywhere in common (#1604) - feat(crashtracking): include `Kind` in crash ping and clarify requirements (#1595) - fix(crashtracking): add process_tags to application field (#1576) - ci: update nightly in CI to 2026-02-08 (#1539) - fix(telemetry)!: fix logs payload format [APMSP-2590] (#1498) - chore(crashtracking): bump os_info crate to 3.14 (#1507) - Don't bail (#1494) - feat(crashtracking)!: emit crashing thread name in crash report for linux crashes (#1485) - test(crashtracking): add minimal LD preload test for crashtracker collector (#1428) - chore: release libddcommon-v1.1.0 (#1456) - chore: prepare libdd-telemetry-v2.0.0 (#1457) - [crashtracker] Retrieve panic message when crashing (#1361) - fix(sidecar): AWS lambda also can return EACCESS for shm_open (#1446) - chore(crashtracking): add `is_crash_debug` tag to crashtracker receiver debug logs (#1445) - [crashtracker] Log errors in crashtracker receiver (#1395) - chore: add changelog for every published crate (#1396) - Fix CI (#1389) - [crashtracker] Set OS info in the crash info builder when receiving report (#1388) - crashtracker: support cxx bindings for crashinfo (#1379) [APMSP-2764]: https://datadoghq.atlassian.net/browse/APMSP-2764?atlOrigin=eyJpIjoiNWRkNTljNzYxNjVmNDY3MDlhMDU5Y2ZhYzA5YTRkZjUiLCJwIjoiZ2l0aHViLWNvbS1KU1cifQ --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: hoolioh <107922352+hoolioh@users.noreply.github.com>
What does this PR do?
Skips one ptrace thread-enumeration test under Miri.
Makes async test assertions wait for the expected request instead of relying on fixed sleeps.
Clears the macOS
target/panic-abortcache before tests as a short-term workaround for stale cached build artifacts.Motivation
These flaky tests and the macOS cache issue are currently blocking builds.
Additional Notes
The macOS cache cleanup is intended as a short-term unblocker. A separate draft PR moves panic-abort artifacts outside the cached target tree.
How to test the change?
cargo +nightly-2026-02-08 fmt --all -- --checkcargo test -q -p libdd-data-pipeline-ffi exporter_send_telemetry_testcargo test -q -p libdd-profiling exporter::exporter_manager::tests -- --nocapturecargo test -q -p libdd-common --features reqwest test_file_dump_captures_http_request -- --nocapturecargo test -q -p libdd-data-pipeline-ffi exporter_send_telemetry_testrepeated 30 timesRUSTUP_TOOLCHAIN=nightly-2026-02-08 MIRIFLAGS="-Zmiri-disable-isolation" cargo miri test -p libdd-crashtracker enumerate_discovers_spawned_thread -- --nocapture