Please refer to the 7.82.1 tag on integrations-core for the list of changes on the Core Checks
Bug Fixes
Windows: Fixed an issue where an explicit DDAGENTUSERKEEPRIGHTS or DDAGENTUSERNAME value passed as an install argument to a Fleet Automation-triggered Windows Agent install/upgrade could be silently overridden by a stale fallback value (res
Fix an issue where GPU monitoring could trigger a kernel panic on multi-GPU nodes with Hopper/Blackwell GPUs.
### What does this PR do? Replaces the hardcoded AIX Python check list with dynamic discovery of every integrations-core check tagged `Supported OS::AIX`. ### Motivation Keep AIX in sync with integrations-core instead of drifting from a hand-maintained list. ### Describe how you validated your changes Ran the updated stage on an AIX 7.3 build host: it instal
### What does this PR do? Wrap `TestKeepTryingLockingIfPermissionDenied`, `TestContextCancellation` and `TestHandleMultipleConcurrentWrites` in `synctest` "bubbles" by extracting their body into `sync<same name>` functions while dropping `t.Parallel()` (disallowed inside a bubble). Drop remaining `t.Parallel()` from the file's other 2 tests too (Bazel schedu
### What does this PR do? First proof-of-concept for LLM Observability built on top of USM/go-tls. When USM observes an HTTP/2 request whose path looks like an LLM API call (e.g. /v1/chat/completions, /v1/messages), we emit one APM span per request (no aggregation) with the full request path as the resource. For connections flagged as LLM traffic, the go-tls
## Summary Temporarily flips \`data_plane.enabled\` from \`false\` to \`true\` in the agent's compiled-in config default so that every e2e test in the full suite runs against ADP without requiring per-test changes or test duplication. The change is throwaway — close without merging once the CI run completes. This is the same config knob used by the AMP ADP t
## Summary This PR contains the reusable delegated-auth foundation for WIF dual shipping: - extends the delegated-auth component so an instance can update an additional-endpoint key in either map- or list-shaped configuration - adds shared helpers for normalizing list-shaped endpoints and reading case-insensitive fields - recognizes pending `DELA(...)` direc
### What does this PR do? Adds a new Windows `powershell` check that runs admin-allowlisted, read-only PowerShell cmdlets and maps their output objects to metrics and tags. The data model mirrors the WMI check — `metrics`, `tag_by`, `tags`, `filters`, and `tag_queries` (joins) — so each cmdlet output object is a "row" and its properties are "columns." Key pi
This is attempt 2 due to needing to revert last version ## Summary - Flips the trace-agent's `convert-traces` v1.0 payload-conversion path to be **enabled by default**. - Introduces a new opt-out feature flag, `disable-convert-traces`, for teams/customers who need to revert to the legacy code path. - Updates `pkg/trace/api/api_test.go` to reflect the new def
## Summary A week-long Error Tracking sweep of installer error categories surfaced this: when the fleet-automation daemon execs a fresh `datadog-installer` child process (for state refresh, install, or experiment operations) and that child crashes at Go-runtime bootstrap because the host is out of memory, threads, or paging capacity — e.g. `fatal error: page
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Enriches the error returned when the Cisco SD-WAN paginated API hits the`max_pages` limit. The error now reports how many pages were fetched, that more data remains, and the current `max_cou
## What changed This adds a testbench-only, timestamp-aware detector view for log-derived occurrence counts. - Keep extractor output sparse in shared Observer storage. - Present completed fixed-window counts and causal forward zeros only when detectors read log `.count` series. - Never backfill zeros before a series is first observed. - Stop virtual zeros af
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? ### Motivation ### Describe how you validated your changes ### Additional Notes
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Populates `cmd/agent/dist/conf.d/network_config_management.d/conf.yaml.example` with the `init_config` and instance-level options supported by the Network Config Management (NCM) check, sour
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Adds **observational-only** eBPF diagnostics to confirm a suspected root cause for encrypted traffic being reported as `tls_encrypted:false` in CNM. **No classification behaviour changes.**
### What does this PR do? Removes the deprecated `Install-Datadog.ps1` script and its build/sign/deploy CI jobs, and migrates the Windows e2e install coverage that depended on it over to `datadog-installer.exe`. ### Motivation The script was already deprecated in favor of `datadog-installer.exe` (see the [deprecation note](https://github.com/DataDog/datadog-
### What does this PR do? Gives the existing `computer_group_policy` / `user_group_policy` milestones a drill-down: the client-side extensions (CSEs) that ran inside each boot Group Policy pass, and the Group Policy objects that fed each one. New `custom.group_policy_details` block, a sibling of the untouched `boot_timeline` and `durations`. This is the real
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? This [commit](https://github.com/torvalds/linux/commit/e67ddd9b1cff7872d43ead73a1403c4e532003d9) landed in kernel v6.9 negatively effects the complexity of the `trace_contention_begin` bpf p
> **Note:** This PR was created by Claude. Stacked on #54652 — review that one first. ### What does this PR do? Adds a `comp/metadata/workloadbalancing` component, mirroring the existing `comp/metadata/haagent`. When Agent workload balancing is enabled it reports a `workload_balancing_metadata` inventory payload: ```json { "hostname": "...", "workload_balanc
### What does this PR do? Extracts a `dd-procmgr-client` Rust crate from client code in `dd-procmgrd`: - process-manager protobuf/gRPC bindings; - default endpoint and `DD_PM_SOCKET_PATH` handling; - Unix socket and Windows named-pipe connections; - Windows busy-pipe retry behavior. The `dd-procmgr` CLI now uses this crate. The daemon reuses its bindings and
### What does this PR do? Adds a new `_prepare_aix_update` step to `tasks/python_version.py` so the `python-version.update` task also bumps `PYTHON_VERSION` in `packaging/aix/lib/env.sh`. Also updates the `upgrade-python-patch-version` workflow's PR description to mention this file, and adds a unit test covering the new function. ### Motivation PR #54696 (th
## Summary This PR adds an `installer_metadata` payload for the fleet installer, collected by the core Agent over a new read-only status socket on the installer daemon. Two pieces: 1. A second listener on the installer daemon serving one route, `GET /status`, returning the installer version and available disk space ([`pkg/fleet/daemon/status_api.go`](https:/
### What does this PR do? Fix build file for cmd/loader so it can also build for macos. - Remove unneeded target_compatible_with - Adjust packaging to include it for macos. - Fix dd_agent_go_binary so we can select() `exact_gotags` We are blocked on ABLD-294 for using the loader in the real package, so we just have it named "loader" for now. We can build and
### What does this PR do? Cuts bazelisk out of tasks/install_tools. Except on mac, because for some reason it is not in the build image. ### Motivation We should not be installing bazelisk from tasks. That muddies the chain to source of trust because bazel is the root of the trusted tools. Bazelisk should come with the build image or be installed in some oth
### What does this PR do? Bumps the AIX embedded Python pin from 3.13.14 to 3.13.15 in `packaging/aix/lib/env.sh`. ### Motivation The AIX packaging build pins the embedded Python version in `packaging/aix/lib/env.sh`, and `packaging/aix/stages/02-python.sh` downloads and builds CPython from that value. This file is **not** touched by `dda inv python-version.
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? `agent.build` task fails unable to find `windmc` tool in case it isn't accessible in `PATH`. This change handles this by letting tasks use `windmc` managed by Bazel. In this case we ensure t
### What does this PR do? It removes the `redacted_compat.h` file out of the sources for the Python build on Windows, such that it doesn't get used for cache computation. ### Motivation I didn't get remote cache hits on a local checkout with git core.autocrlf disabled (see also https://github.com/DataDog/datadog-agent/pull/54776) due to this specific file. S
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Replaces the Fleet Automation process/service check in core Agent status with the installer local API. The local API wire contract and client now live in the updater components instead of `p
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? ### Motivation ### Describe how you validated your changes ### Additional Notes
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? ### Motivation ### Describe how you validated your changes ### Additional Notes
Backport d1520bca29c837ff691f91102b60a32a37c0112c from #54563. ___ ### What does this PR do? Gates the NVML workloadmeta collector on `gpu.enabled`. ### Motivation Avoid NVML collection when GPU monitoring is disabled. ### Describe how you validated your changes Added unit coverage for disabled startup. ### Additional Notes
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Moves tar_checksums code to under `bazel/tools`, such that it falls under the agent-build team ownership. ### Motivation I made a PR that touched these files and no review from agent-build w
Backport e5a19fa15faff217145de62c45995fd1f2451ee1 from #54729. ___ <!-- dd-meta {"pullId":"00000000-0000-0000-0000-000000000000","source":"chat","resourceId":"b6950850-80dc-4801-bc2f-895d5b466cca","workflowId":"c63cfe10-11dd-4bbc-800e-23188cad9c50","codeChangeId":"c63cfe10-11dd-4bbc-800e-23188cad9c50","sourceType":"bits_ai_sre"} --> ### What does this PR do?
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Builds are notoriously slow on Windows so it was requested by @DataDog/windows-products to fail the agent build in case we can't utilize the cache. One of the main reasons for that is absenc
### What does this PR do? `TotalSeriesCount` was taking a lot of CPU time (and mostly for telemetry), removed overhead by O(1) CPU / memory. This removes 36.8% of CPU on the staging cluster I tested, quick win! ### Motivation ### Describe how you validated your changes [Profile](https://ddstaging.datadoghq.com/profiling/comparison?query=service%3Adatadog-age
### What does this PR do? Mutes `test/new-e2e/tests/agent-subcommands.TestWindowsFlareSuite/TestFlareDefaultFiles` in `flakes.yaml`. ### Motivation This test has been reported failing intermittently in CI, tracked in [FLREM-146](https://datadoghq.atlassian.net/browse/FLREM-146). `TestWindowsFlareSuite` is a testify suite with other subtests that were never r
### What does this PR do? Adds group- and version-aware resolved workload targets to Kubernetes Pod workload metadata and the workload-filter CEL model. - Preserves `apiVersion` and `controller` from Kubernetes owner references in both Pod parsers. - Adds `KubernetesPod.ResolvedTargets` with group, version, kind, namespace, name, and UID identity. - Exposes
Backport 5dcbd3081dd283e30cf25879eb5be7b55577e50a from #54720. ___ <!-- dd-meta {"pullId":"df13230b-a434-451b-972e-ac9007c02168","source":"chat","resourceId":"86800824-17f2-4a85-9551-be5b7bf8a830","workflowId":"2ce581a4-a03e-4c0a-a1a1-dda5b27f5a0c","codeChangeId":"2ce581a4-a03e-4c0a-a1a1-dda5b27f5a0c","sourceType":"bits_ai_sre"} --> ### What does this PR do?
## TL;DR Fixes a nil pointer dereference in the `cloudfoundry-vm` workloadmeta collector that crashes the node Agent on Cloud Foundry when the connection to the Cluster Agent fails. Regressed in 7.81. ## Motivation Agents on Cloud Foundry running 7.81+ panic repeatedly with: ``` panic: runtime error: invalid memory address or nil pointer dereference clustera
What does this PR do? Fixes Network Path dynamic tests scheduling traceroutes against Datadog's own intake backends (e.g. l4-metrics-agent-*.elb.us-east-1.amazonaws.com) despite the default *.datadoghq.com connection-filter exclude. - Adds ConnFilter.EvaluateDomains(domains, ip), which evaluates every DNS name a destination IP reverse-resolves to (via the ex
## Context PRs #54652, #54656, #54659, and #54795 implemented NDM Agent Workload Balancing's agent-side component (`comp/workloadbalancing`), registering its own Remote Config product, `NDM_AGENT_WORKLOAD_BALANCING`. The design RFC ([NDM Agent Workload Balancing: Device Handoff](https://datadoghq.atlassian.net/wiki/spaces/II/pages/7029621750)) has since been
## Summary This adds the environment-variable-driven wiring code for the MicroVM lifecycle HTTP server: reading `DD_AWS_MICROVM_LIFECYCLE_PORT` and the other lifecycle-related env vars, and assembling the config needed to construct the server. No HTTP server or handlers are introduced yet — those land in the next PR in this stack. The motivation is to let th
### What does this PR do? Adds a `*ProcessHooks` parameter (`OnAlive`/`OnDead`) to `mode.RunInit` and its internal `execute()`, so a caller can be notified when the spawned user process starts and exits. `mode.Conf.Runner` is dropped for init-container mode — its field type (`func(*serverlessLog.Config) error`) can no longer match `RunInit`'s new two-argumen
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? `dd-compile-policy` binary is pinned within the Agent at v0.1.2, which predates https://github.com/DataDog/dd-policy-engine/pull/68 in `dd-policy-engine` that added Docker container evaluato
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? This PR adds the `kube_cronjob` tag to Kubernetes Job events when they are spawned by a cronjob. It does this by parsing through the job name, the same as what is done in the [tagger](https:
### What does this PR do? Collects selected, non-secret Redis environment variables alongside Redis config files in `configfilesdiscovery`. It uses regex-based allow and deny rules plus the shared secret-name filter. Config-file selection follows `redis-server` argv, then `REDIS_CONF_FILE`, then known default paths. The shared reader accepts the optional fal
### What does this PR do? Packages and activates `par-control` on Linux and Windows: - Adds Agent package and installer wiring. - Installs process-manager definitions for `par-control` and the executor. - Adds Windows executable resources and MSI integration. - Preserves a standard service `PATH` for Unix process-manager children. - Adds Linux and Windows sp
<!--Please give us some feedback on your experience writing this PR ! https://app.datadoghq.com/forms/43db4c02-6837-400c-8083-692e141b1b88 !--> ### What does this PR do? Adds a new metric on the agent, `datadog.ncm.check_failure`, which counts NCM config-check failures. ### Motivation [Inspiring Slack thread](https://dd.slack.com/archives/C08NTAS3A1E/p178585