fix(shared-mc): retry stranded HEC batches instead of killing sender thread - #26
Open
shiftySenpai wants to merge 7 commits into
Open
fix(shared-mc): retry stranded HEC batches instead of killing sender thread#26shiftySenpai wants to merge 7 commits into
shiftySenpai wants to merge 7 commits into
Conversation
Extract buildHecEnvelope() helper in SingleSplunkConnection using gson
(already a shared-mc dependency) instead of the unmaintained json-simple
1.1 library. Wire format is unchanged: {"event": <message>}. Adds unit
tests covering basic wrapping and quote escaping, and removes the
json-simple dependency from shared-mc/pom.xml and the parent pom's
dependencyManagement.
…'d deps Adds per-platform MC-version Maven profiles (spigot: 1.21.1 default plus 1.20.1/1.20.4/1.20.6), a shade-plugin build producing logtosplunk-<mc.version>-<loader>-<loader.version>.jar per module, and marks server-provided deps (log4j, spigot-api) as provided scope so they aren't shaded into the plugin jar. Pins kotlin-stdlib to 2.0.21 to clear CVE-2026-53914 pulled in transitively via splunk-library-javalogging -> okhttp3 -> kotlin-stdlib:1.6.20. Tunes the OWASP dependency-check plugin to skip provided-scope CVEs (server operator's responsibility) and not fail on Sonatype OSS Index 401s (paid-only), while still enforcing the CVSS 7.0 gate.
…der thread Two related failure modes in SingleSplunkConnection's send loop, both of which permanently stopped all Splunk logging until the server restarted: - If sendData() threw for any reason, the sender thread's run() loop wasn't catching it, so the whole thread died silently on the first transient error (e.g. HEC briefly unreachable at startup). - If an HTTP execute() call failed, response was never assigned, but the finally block unconditionally dereferenced it to close it -- turning every send failure into an uncaught NPE. The failed batch was also left on the "runway" and, due to an early-return bug, never retried. Now: send failures are caught and logged per-cycle (thread keeps running), a failed batch is retried next cycle instead of being stranded, and both response/httpClient closes are null-guarded.
6 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Split out of #23 (reliability-fix slice). Stacked on #24 (build/deps) — this branch includes #24's commits, so the diff below will look large until #24 merges, after which GitHub will automatically shrink it to just this PR's commit. Independent of the extended-logging PR.
SingleSplunkConnection's HEC sender thread used to die permanently on a single send failure, requiring a server restart to recover. Failures are now caught and the batch is retried instead of killing the thread.Test plan
mvn clean package(not verified in this environment — no local Maven install; please build-check before merging)Part of the #23 split.