APACHE SPARK COVERAGE
Spark 3.4 and 3.3 are end of life. Your patching does not have to be.
Apache Spark extended support from OSSeva keeps Spark 3.4 and 3.3 patched after the Spark project ended them: 3.4.4, the last 3.4 release, shipped on 27 October 2024 and 3.3.4 on 16 December 2023. Critical CVEs are fixed within 48 hours and High within 7 days, priced per application. Other versions are assessed case by case through a version coverage check.
| Who it is for | Data engineering teams running Spark 3.3 or 3.4 on YARN, Kubernetes or standalone clusters that cannot move to Spark 4 yet |
|---|---|
| What you get | Signed patched builds on the line you run, with Java, Scala and Python versions and your job code unchanged. Assure adds a Spark 4 upgrade plan. |
| How it is priced | Per application. Ask for a quote. |
| What happens next | You send your work email, we confirm your Spark line, Scala build and cluster manager, and you get a coverage plan. |
Reply within 1 business day · No commitment · or book a 15-minute call
Last reviewed
Trusted globally by enterprises





Apache Spark support status by line, read on 19 November 2026
Release dates and latest releases are from spark.apache.org. End of life is from endoflife.date, except 3.5, where the project states November 2027 itself.
| Line | First release | End of life | Latest release listed |
|---|---|---|---|
| 4.1 | 16 Dec 2025 | 11 Jun 2027 | 4.1.4 (27 Oct 2026) |
| 4.0 | 23 May 2025 | 23 Nov 2026 | 4.0.5 (27 Oct 2026) |
| 3.5 LTS | 13 Sep 2023 | 30 Nov 2027, security fixes only | 3.5.9 (16 Jul 2026) |
| 3.4 | 13 Apr 2023 | Oct 2024 | 3.4.4 (27 Oct 2024) |
| 3.3 | 16 Jun 2022 | Dec 2023 | 3.3.4 (16 Dec 2023) |
Spark 4.2 and 4.3 are also released; their dates are on the end-of-life page.
Spark versioning policy →Spark release news →endoflife.date Spark data →Apache Spark end-of-life dates →
What is in the way of moving Spark 3.3 or 3.4 forward
Spark 3.3 and 3.4 clusters stay where they are for build reasons more often than for lack of effort. The release notes and documentation show where.
- Scala is the first constraint. Spark 3.5 supports Scala 2.12 and 2.13; Spark 4.0 dropped 2.12, so every internal library and connector built as _2.12 needs a _2.13 build before the job can move.
- Java is the second. Spark 3.5 runs on Java 8, 11 or 17, while Spark 4.0 dropped JDK 8 and 11 and made JDK 17 the default. The 4.3.0 documentation lists Java 17, 21 or 25.
- PySpark has its own floor. Spark 4.0 dropped Python 3.8, raised the minimum pyarrow to 11.0.0 and NumPy to 1.21, and upgraded Pandas to 2. The 4.3.0 documentation says Python 3.11 or later, so a PySpark estate on an older interpreter changes that too.
- SQL results can change. Spark 4.0 turns ANSI SQL mode on by default, so lenient casts and silent overflow in old queries now fail or return different values. Compare outputs on production data before cutover.
- Mesos is gone. Spark 4.0 removed Mesos support, so Mesos-managed clusters move to YARN, Kubernetes or standalone first.
- Security fixes do not always reach old lines. CVE-2025-55039 is fixed in 3.4.4, 3.5.2 and 4.0.0 and affects every version before them, so a cluster on 3.3, whose last release is 3.3.4, has no upstream fix. It applies only when spark.network.crypto.enabled is true.
Sources: the Spark 4.0.0 release notes, the 3.5.9 and 4.3.0 documentation, spark.apache.org/security.html and the CVE-2025-55039 record on cveawg.mitre.org, which agree on the affected versions, read on 19 November 2026.
Hadoop support →Hive support →Flink support →Kafka support →Python versions and end dates →
Why now
Spark 3.3 and 3.4 have no further releases
The last Spark 3.3 release was 3.3.4 on 16 December 2023 and the last 3.4 release was 3.4.4 on 27 October 2024, per spark.apache.org. endoflife.date records their end of life as December 2023 and 21 October 2024.
A fix can exist for 3.4 and not for 3.3
CVE-2025-55039, on unauthenticated AES-CTR RPC encryption, affects every Spark version before 3.4.4, 3.5.2 and 4.0.0. No 3.3 release carries the fix, because 3.3.4 came out before it.
Spark 4 changes the platform under every job
Spark 4.0 dropped JDK 8 and 11, dropped Scala 2.12, dropped Python 3.8 and turned ANSI SQL mode on by default. Moving from 3.x is a rebuild of every _2.12 dependency and a result comparison, not a version bump.
Versions covered
All versions below receive active CVE patches from OSSeva. Version numbers in monospace are exact release identifiers.
| Version | Status | Active CVEs |
|---|---|---|
| 3.3(Community EOL Dec 2023 (last release 3.3.4)) | EOL | See advisories |
| 3.4(Community EOL Oct 2024 (last release 3.4.4)) | EOL | See advisories |
| 3.5(LTS, security fixes only to Nov 2027. Assessed case by case.) | Extended | See advisories |
| 4.0(Community EOL 23 Nov 2026. Assessed case by case.) | Current | See advisories |
| 4.1(Community EOL Jun 2027. Assessed case by case.) | Current | See advisories |
| 4.2(Community EOL Jan 2028 per endoflife.date. Assessed case by case.) | Current | See advisories |
| 4.3(Latest release, 4.3.0 on 30 Oct 2026. Assessed case by case.) | Current | See advisories |
Running a Apache Spark version or build not listed here?
What you get
Three tiers — pick the level of engagement that matches your team's operational needs and compliance requirements.
OSSeva Patch
Security patches for Apache Spark 3.4 and 3.3 after end of life.
- Backported security fixes for Apache Spark 3.4 and 3.3
- Critical CVEs fixed within 48 hours, High within 7 days
- Signed artifacts (GPG + Sigstore)
- CVE disclosure notifications
- Architecture review
- 24/7 managed operations
OSSeva Assure
Patch plus compliance documentation and a Spark 4 upgrade plan.
- Everything in Patch
- Inventory of Spark versions, Scala builds and cluster managers across jobs
- Upgrade path assessment from 3.x to Spark 4
- SOC 2 / HIPAA attestation package
- 24/7 reactive support
OSSeva Operate
Full MSP: 24/7 monitoring, 15-min P1 response, named engineers.
- Everything in Assure
- 24/7 cluster and job monitoring
- 15-minute P1 incident response SLA
- Named senior engineer on your account
All tiers priced per cluster/application — not per core.
Which Apache Spark versions does OSSeva patch?
Apache Spark 3.4 and 3.3 after end of life. Other versions are assessed case by case through a version coverage check.
Is Apache Spark 3.4 end of life?
Yes. endoflife.date records October 2024 as its end of life, and 3.4.4, listed on spark.apache.org on 27 October 2024, was the last 3.4 release.
When does Apache Spark 4.0 reach end of life?
On 23 November 2026, per endoflife.date. The Spark versioning policy does not state a date for 4.0.
Is Spark 3.5 supported?
Yes, for now. The Spark versioning policy gives 3.5.x an extended LTS period to November 2027 with security fixes only. endoflife.date dates it 30 November 2027, and the last release read was 3.5.9 on 16 July 2026.
How fast are Spark vulnerabilities patched?
Critical vulnerabilities (CVSS 9.0 or higher) within 48 hours and High within 7 days. Everything else goes into a scheduled batch.
How is OSSeva Apache Spark support priced?
It is priced per application. Request a quote and a senior engineer replies within 1 business day.
Get a quote
Ready to get Apache Spark patched and supported?
Tell us your Apache Spark versions and an engineer replies within 1 business day with coverage and a price. Prefer a call? It takes 15 minutes.
Prefer to talk or read more first?