Raw dependency graph data (JSON)

InfluxDB

Author: Ludovic HENRY ludovic.henry@qti.qualcomm.com
Date: 2026-09-07
Readiness: orange
Scope: RISC-V (riscv64/linux) support status for InfluxDB
Audience: Technical leadership, resource allocation strategy
Verification policy: Every claim is cross-referenced to a primary upstream source. Items that could not be verified against a second source are marked [NEEDS VERIFICATION].

1. Project Overview

InfluxDB is a time-series database. The repository at influxdata/influxdb currently defaults to InfluxDB 3 Core, a complete Rust rewrite (crates under influxdb3_*, workspace Cargo.toml), superseding the older Go-based 1.x (master-1.x branch) and 2.x (main-2.x branch, using go.etcd.io/bbolt) codebases.

Governance: single-vendor, corporate-controlled. There is no neutral foundation (not CNCF, not Apache Software Foundation, not Linux Foundation). The project is owned and directed by InfluxData, Inc. No MAINTAINERS, OWNERS, or GOVERNANCE.md file exists in the repo; .github/CODEOWNERS contains only boilerplate comments with no configured ownership rules. CONTRIBUTING.md requires contributors to sign InfluxData’s Individual CLA (influxdata.com/legal/cla), and InfluxData staff unilaterally triage and decide what gets merged.

License: dual-licensed MIT / Apache-2.0 (permissive), per LICENSE-MIT and LICENSE-APACHE.

Top contributors (from git shortlog), all apparently InfluxData employees: Trevor Hilton (93 commits), praveen-influx (39), wayne (29), Phil Bracikowski (26), Michael Gattozzi (24), peterbarnett03 (20), Jamie Strandboge (19), Stuart Carnie (17), Paul Dix (6, InfluxData CEO/founder). No external-company maintainers were identifiable.

Community culture on new ports: reactive, not structured. There is no public RFC or tiered-support process, and no PLATFORMS.md/SUPPORT.md policy document exists. Feature/platform requests are periodically reviewed and accepted or rejected at InfluxData’s discretion. There is no evidence of active hostility toward a riscv64 port, but also no evidence of any interest expressed by a maintainer.

2. Port History and Upstreaming Timeline

Date Event Source
2019-08-28 InfluxDB 2.x migrates from boltdb/bolt import path to bbolt (named imports) PR #14841
2019-09-18 InfluxDB 2.x fully switches to etcd-io/bbolt as the storage dependency PR #13917
2021-01-20 Issue opened proposing “Replace boltdb/bolt with etcd-io/bbolt” – filed after the migration above had already happened; issue remains open/unclosed to date, effectively orphaned Issue #20559
2022-09-03 Third-party Alpine Linux packaging bug report is filed; the pasted APKBUILD snippet contains arch="all !riscv64" # riscv64 blocked by boltdb dependency – the only riscv64 reference ever found in the repo’s issue history, and it is downstream packaging metadata, not an InfluxData statement Issue #23709
2026-09-07 Verification pass confirms go.mod on main-2.x already depends on go.etcd.io/bbolt v1.3.6, not boltdb/bolt – the boltdb blocker claim quoted in #23709 is stale/incorrect as of today Direct repo inspection, main-2.x go.mod, this research pass

Key contributors to any riscv64 work: none. No individual or organization (InfluxData or external) has ever authored a riscv64-related commit, PR, or tracking issue in this repository. search_pull_requests, search_commits, and search_code for “riscv”/”riscv64” against influxdata/influxdb all return zero results.

Is it fully upstream? No. There is no riscv64 code, CI job, or release asset upstream at all. The single downstream artifact that exists (an Ubuntu .deb, see Section 8) is not from InfluxData and is not a current build.

3. Upstream Support Tier

No formal tier policy exists (no SUPPORT.md/PLATFORMS.md). In practice, support tiers are defined implicitly by the CircleCI build/release matrix, which InfluxData treats as release-blocking for the architectures it lists.

Architecture CI builds CI tests Official release binaries Notes
amd64/x86_64 Yes Yes Yes (linux, darwin, windows, .deb, .rpm) Primary target
arm64/aarch64 Yes Yes Yes (linux, darwin, .deb aarch64, .rpm aarch64) Fully supported, identical release cadence to amd64
riscv64 No No No Absent from CI matrix and from every recent release’s asset list

Source: .circleci/config.yml (1005 lines) and .circleci/packages/config.yaml read directly at HEAD 693b1fd; target list is exactly x86_64-unknown-linux-gnu, aarch64-unknown-linux-gnu, aarch64-apple-darwin, x86_64-pc-windows-gnu/msvc. Release asset confirmation via the GitHub releases page (v3.10.0, v2.9.1, v2.9.0, v1.12.4, v3.9.0 all checked).

4. Technical Architecture and RISC-V-Specific Subsystems

InfluxDB has no hand-written, per-architecture source files for any architecture, not just riscv64. Code search for GOARCH, per-arch Go build-tag filenames (_amd64.go, _arm64.go), and assembly (extension:s) against influxdata/influxdb all returned zero hits. Multi-architecture support in this project exists entirely at the build/release/CI layer (cross-compiled binaries and a CI asset matrix), not as arch-specific source code.

The current default branch (InfluxDB 3, Rust) does carry architecture-sensitive dependencies at the crate level, most notably a WASM JIT for user-defined-function hosting:

Component amd64 arm64 riscv64
Query execution (DataFusion/Arrow, scalar+SIMD kernels) Full Full Builds via portable scalar fallback; no known blocker (apache/arrow-rs, apache/datafusion issue search returns zero riscv64 hits)
Crypto (ring, under rustls) Full Full Supported; riscv64gc issues in briansmith/ring are closed/resolved (#2148, #2468, #2520)
Hashing (BLAKE3) Hand-tuned SIMD Hand-tuned SIMD Portable/scalar path only; no vector SIMD codepath yet, tracked as a performance (not correctness) gap (BLAKE3-team/BLAKE3 #484)
WASM JIT for UDF hosting (wasmtime/cranelift-codegen) Mature Mature Immature and reported broken on real hardware. Multiple open, unresolved codegen crashes: wasmtime #13959 (Cranelift crash in riscv64 isle, repros on real Gentoo riscv64gc hardware), #12195 (ISLE panic in gen_bitcast), #13078 (vxrm/vxsat registers not preserved – a vector-extension correctness bug), #9186 (*_overflow unsupported on riscv64)
Allocator (jemalloc) Full Full Packaged and functional (Ubuntu 26.04 riscv64 ships librust-tikv-jemalloc-sys-dev etc.); a musl-specific gap exists (tikv/jemallocator #118) but does not affect the glibc riscv64 build

InfluxDB is not itself an optimization-purpose project (it is a database server/runtime, not a SIMD/allocator/compression library whose sole value proposition is raw speed over a naive baseline), so the Step 2 optimization-cap methodology does not apply to InfluxDB as a whole. The wasmtime/Cranelift risk above is a functional/correctness concern for one feature (WASM-hosted UDFs), not a performance-optimization gap.

5. Build System, Cross-Compilation, and Toolchain

InfluxDB 3 (current main) is a pure Rust/Cargo project. There is no CMake anywhere in the repository (CMakeLists.txt and *.cmake both absent).

  • Toolchain: Rust, pinned via rust-toolchain.toml to channel = "1.97.1" (components: rustfmt, clippy, rust-analyzer, rust-src). No project-specific minimum GCC/Clang version is documented; the project does not compile C/C++ code itself beyond what Rust dependency build scripts invoke via the system cc/clang.
  • Build commands (from CONTRIBUTING.md): 
    git clone https://github.com/influxdata/influxdb.git
cd influxdb
cargo build                      # default (unoptimized) profile
cargo build --profile release    # optimized release build
cargo build --profile quick-release
cargo build --profile quick-bench
  • System dependencies: a working python3 install and the protoc Protocol Buffers compiler.
  • Docker build (root Dockerfile): base image rust:1.92-slim-bookworm, installs binutils build-essential curl pkg-config libssl-dev clang lld git patchelf protobuf-compiler zstd libz-dev, then runs cargo build --target-dir /influxdb3/target --package="$PACKAGE" --profile="$PROFILE" --no-default-features --features="$FEATURES" with default FEATURES=aws,gcp,azure,jemalloc_replacing_malloc.
  • QEMU usage: none found anywhere in the repository.
  • Known build failures relevant to non-x86/arm64 architectures: Issue #22585 (open) reports the older Go+Rust libflux component failing to link on mips64le with undefined reference to 'fmaf'/'fma' – a Rust-std math-intrinsic/libm linkage gap on an uncommon architecture. This is not riscv64 itself but illustrates the same class of risk for any Rust cdylib boundary on less-common targets. No riscv64-specific build failure has ever been reported, because no one has attempted the build (see Section 11).
  • No riscv64 build documentation exists. BUILDING.md, docs/building.md, docs/cross-compilation.md, cmake/riscv64.cmake, and any docker/Dockerfile.riscv64 do not exist in the repository. A repository-wide, case-insensitive grep for “riscv” across all file types returns zero matches. Anyone building for riscv64 today would need to do so unofficially via plain Rust cross-compilation (rustup target add riscv64gc-unknown-linux-gnu plus a suitable linker), entirely undocumented and unsupported by upstream.

6. Feature Coverage and Gap Analysis vs arm64 and amd64

Dimension amd64 arm64 riscv64
Compiles from upstream source Yes Yes Untested by upstream; no upstream CI evidence either way
Official release binary Yes Yes No
Official Docker image Yes Yes No
WASM-hosted UDF feature (processing engine) Functional Functional High risk: wasmtime/Cranelift has open, unresolved riscv64 codegen crash bugs on real hardware (Section 4)
Hashing performance (BLAKE3) SIMD-accelerated SIMD-accelerated Scalar-only fallback; expected to be slower, though no InfluxDB-specific benchmark exists to quantify the delta (Section 8/11)
Security hardening posture N/A – no riscv64-specific finding N/A No data found either way; not evaluated by any source in this research
NaN / floating-point semantics N/A – no reported issue N/A No riscv64-specific floating-point issue found; note that the only open non-x86 architecture issue in the tracker is #27278, an ARM64-vs-x86_64 performance comparison (not riscv64, and not FP-semantics related)

Functional gap: the WASM UDF-hosting feature is the one concretely identified functional risk area for riscv64, rooted in upstream wasmtime/Cranelift immaturity on that architecture, not in InfluxDB’s own code. Performance gap: cannot be quantified – no InfluxDB riscv64 benchmark of any kind exists (Section 11). Security-hardening and NaN/floating-point gaps: no data found; not claimed either way.

7. CI/CD Infrastructure

No riscv64 CI exists, confirmed by direct reads of every CI configuration file present in the repository at HEAD 693b1fd1b96cdcb980cf76a1004c0b3f1b46db48 (2026-09-07):

  • .github/workflows/semantic.yml – the only GitHub Actions workflow present. It is a PR-title/commit-message linter (uses: influxdata/validate-semantic-github-messages/.github/workflows/semantic.yml@main), triggered on pull_request: [opened, reopened, synchronize, edited]. It has no build or test jobs and no runner matrix at all.
  • .circleci/config.yml (1005 lines) and core/.circleci/config.yml (151 lines) – InfluxDB’s actual build/release CI. A direct grep (grep -in "riscv" .circleci/config.yml .circleci/packages/config.yaml ...) returns zero matches. Extended search for variants (risc-v, riscv64, riscv32, rv64, rv32gc) likewise returns zero matches.

No RISE runner references (riseproject-dev, RISE runner labels) appear anywhere in either CI system. No hardware of any kind is used for riscv64 because no riscv64 job exists.

Architecture CI system Build job Test job Release-blocking Runner
amd64 CircleCI Yes (x86_64-unknown-linux-gnu, plus package validation check_package_deb_amd64/check_package_rpm_amd64 via docker run --platform linux/amd64) Yes Yes CircleCI executor
arm64 CircleCI Yes (aarch64-unknown-linux-gnu, plus check_package_deb_arm64/check_package_rpm_arm64 via --platform linux/arm64) Yes Yes CircleCI executor
riscv64 None No No N/A N/A

8. Distribution and Release Status

No official riscv64 binaries from InfluxData. The last five GitHub releases (v3.10.0, v2.9.1, v2.9.0, v1.12.4, v3.9.0) ship linux amd64/arm64, darwin amd64, windows amd64, .deb (amd64/arm64), .rpm (x86_64/aarch64) only – no asset filename in any of them contains “riscv”. Source: influxdata/influxdb releases.

PyPI (influxdb Python client): current release 5.3.2, shipped only as a universal wheel (influxdb-5.3.2-py2.py3-none-any.whl) plus sdist. This is a pure-Python client with no compiled extensions, so it is architecture-independent and installs on riscv64 by construction; it does not represent a riscv64 build of the InfluxDB server. Full simple-index history (0.1.1 through 5.3.2) contains no riscv-specific artifact, as expected. Source: pypi.org/pypi/influxdb/json.

RISE Python wheel builder: https://gitlab.com/api/v4/projects/56254198/packages/pypi/simple/influxdb/ returns an HTTP 302 redirect straight to upstream PyPI – RISE has not built a separate wheel for this package (unsurprising, since the PyPI package is already architecture-independent).

Arch Linux RISC-V (archriscv.felixc.at): no influxdb package listed at all.

Ubuntu 26.04 (resolute): the only riscv64 binary package of the InfluxDB server that exists anywhere. Confirmed via packages.ubuntu.com and the direct riscv64 build page (packages.ubuntu.com/resolute/riscv64/influxdb):

  • Package influxdb, version 1.6.7~rc0-2build2, classified [ports] [universe], built by the Debian Go Packaging Team / Ubuntu MOTU Developers – a third-party distro repackaging, not an InfluxData artifact.
  • This version is roughly eight years old, an InfluxDB 1.6 release candidate that predates even the last official 1.x release (1.12.4) cited above and predates the 2019 boltdb-to-bbolt migration context entirely. It has no Flux, no 2.x/3.x API, and does not represent the software currently developed at influxdata/influxdb in any meaningful sense.
  • The identical 1.6.7~rc0 version string has been carried unchanged (only Debian-revision bumps, e.g. -2build2) across jammy (22.04), noble (24.04), questing, and resolute (26.04) – consistent with an unmaintained package riding along automatically rather than reflecting deliberate riscv64 enablement or re-verification.
  • No evidence was found that this riscv64 build has been confirmed to run correctly. A structurally analogous boltdb runtime failure was documented on s390x in Issue #22968 (“unable to open boltdb file invalid database on s390x”), indicating boltdb’s byte-order/format sensitivity is a real risk class on uncommon architectures; whether it affects this specific riscv64 build is unverified either way.
  • python3-influxdb and python3-influxdb-client are packaged [all] (architecture-independent) – installable on riscv64 by nature of being noarch, but these are client libraries, not the server.

What a user must do today to get a working riscv64 InfluxDB binary: either (a) install the stale, third-party-packaged, eight-year-old Ubuntu influxdb 1.6.7~rc0 package with unverified riscv64 runtime correctness, which is not the actively developed codebase, or (b) build InfluxDB 3 from source via unofficial, undocumented Rust cross-compilation (rustup target add riscv64gc-unknown-linux-gnu), with no guarantee that all features (notably WASM-hosted UDFs, Section 4) will work correctly. There is no supported path to a current, InfluxData-built riscv64 binary.

9. Dependencies

Dependency Role riscv64 build riscv64 test riscv64 release Community/notes
wasmtime / cranelift-codegen WASM JIT for UDF hosting Not packaged in Ubuntu 26.04 riscv64; buildable from source but immature backend Reported broken on real hardware (mid-2026) Not distro-packaged for any arch 108 riscv64 issues total in bytecodealliance/wasmtime, most open. Highest-risk dependency identified in this research.
rustls (+ webpki, native-certs, platform-verifier) Pure-Rust TLS Not packaged in Ubuntu 26.04 riscv64 (packaging gap only) N/A N/A Zero riscv64 issues in rustls/rustls; appears to be an archive-side packaging gap, not a code blocker – builds from source via cargo
ring 0.17.14 Crypto backend under rustls Packaged in Ubuntu 26.04 riscv64 (librust-ring-dev) – Packaged riscv64gc issues #2148, #2468, #2520 all closed/resolved
Apache Arrow (arrow-rs) 57.3.0 Columnar format + SIMD compute kernels Not distro-packaged for any arch (vendored via cargo) – – Zero riscv64-specific bugs found; portable scalar fallback covers riscv64
DataFusion 51.0.0 (InfluxData fork) SQL/query execution engine Not distro-packaged – – Zero riscv64 issues in apache/datafusion
tikv-jemalloc (sys/ctl/allocator) Global allocator, heap profiling Packaged in Ubuntu 26.04 riscv64 Likely tested via glibc CI Packaged tikv/jemallocator #118 is musl-only, does not affect Ubuntu’s glibc riscv64 build
BLAKE3 1.8.5 Content hashing (catalog/cache layer) Packaged in Ubuntu 26.04 riscv64 Runs correctly on real riscv64 hardware (Orange Pi RV2 / SpacemiT X60), scalar path only Packaged BLAKE3-team/BLAKE3 #484 (open): performance-only gap, no vector SIMD codepath yet, roughly 3-5x slower on xof benchmarks vs an accelerated architecture
zstd 0.13.3 binding Parquet/object_store compression Packaged in Ubuntu 26.04 riscv64 (libzstd-dev, libzstd1) – Packaged None found
brotli 8.0.2 (pure-Rust) Parquet/object_store compression Packaged in Ubuntu 26.04 riscv64 – Packaged None found
snap / crc32fast / flate2 Additional compression codecs Packaged in Ubuntu 26.04 riscv64 – Packaged None found
PyO3 0.29 Embeds CPython for the Python processing-engine plugin Packaged in Ubuntu 26.04 riscv64 (librust-pyo3-dev, python3-dev) – Packaged Only 1 issue hit in PyO3/pyo3 search and it is s390x-related; zero riscv64 issues
SQLite / sqlx (libsqlite3-sys 0.30.1, sqlx 0.8.6) Embedded metadata store Packaged in Ubuntu 26.04 riscv64 – Packaged None found

Deep dive – wasmtime/Cranelift (highest-risk dependency): this is the WASM JIT hosting InfluxDB 3’s processing-engine user-defined functions (datafusion-udf-wasm-host/-query crates). Ubuntu 26.04 does not package wasmtime/librust-wasmtime-dev/librust-cranelift-dev for riscv64 at all (only amd64/arm64), and upstream bytecodealliance/wasmtime has multiple open, unresolved riscv64 codegen bugs confirmed on real hardware as recently as July 2026, including a vector-extension register-preservation correctness bug (#13078). This is the dependency most likely to require a workaround (e.g., disabling the WASM UDF feature) for any riscv64 InfluxDB 3.x build attempt.

11. Known Bugs and Active Issues

ID Title Status Severity Notes
#20559 Replace boltdb/bolt with etcd-io/bbolt Open (orphaned) Low (stale) Migration already completed via PR #13917/PR #14841 in 2019; issue never closed. Not a real current blocker.
#23709 Build Fails in Alpine Linux Closed N/A Musl static-linking bug (libflux vs musl libc symbol collisions), unrelated to riscv64 except for a quoted, now-stale third-party APKBUILD comment
#22968 Unable to open boltdb file invalid database on s390x Closed Correctness Demonstrates boltdb’s format/byte-order fragility on uncommon architectures – same dependency class relevant to any legacy-1.x riscv64 build, though not riscv64 itself
#22585 Build fail on loongson (mips64le): undefined reference to fmaf/fma Open Build Rust-std libm linkage gap on an uncommon architecture; same risk class could affect riscv64 cross-compilation of Rust cdylib boundaries, unverified
#27278 [v2] Performance comparison between ARM64 and x86_64 for InfluxDB Open Info Not RISC-V; the only open non-x86 architecture performance issue in the tracker, cited here to show the absence of any equivalent riscv64 issue

No riscv64-specific issue, PR, or commit exists anywhere in influxdata/influxdb – confirmed by search_issues (“riscv” 0, “riscv64” 5 results with none textually relevant except the quoted APKBUILD in #23709, “risc-v” 0), search_pull_requests (“riscv” 0, “riscv64” 0, “boltdb riscv” 0), search_commits (“riscv” 0, “riscv64” 0), and search_code (“riscv” 0, “riscv64” 0). There is consequently no correctness bug, performance bug, or benchmark to highlight for riscv64 specifically – the absence reflects lack of any attempted riscv64 usage, not a confirmed-working state.

12. Objections and Upstream Blockers

Stated objections: none found. No InfluxData maintainer has commented on riscv64 anywhere in the issue tracker (the sole quoted riscv64 reference, in #23709, drew zero maintainer response).

Technical blockers:

  • Primary: riscv64 was simply never added to the CircleCI build/release matrix – this is a build-matrix gap, not a code-level blocker, as confirmed by direct reading of .circleci/config.yml.
  • Secondary/real but narrower: the wasmtime/Cranelift WASM JIT dependency used for the processing-engine’s WASM UDF hosting has open, unresolved upstream riscv64 codegen bugs (Section 9), which would need to be resolved or worked around (e.g., feature-gated off) for a full-featured riscv64 build of InfluxDB 3.
  • The historically cited blocker (“riscv64 blocked by boltdb dependency”, per the Alpine APKBUILD note quoted in #23709) is verified stale: InfluxDB 2.x replaced boltdb/bolt with etcd-io/bbolt in 2019, years before this note was written, and the current default-branch codebase (InfluxDB 3, Rust) contains no boltdb dependency at all.

Organizational blockers: InfluxData is a single-vendor, corporate-governed project with no external contributor community structure and no formal process for adding a new CI target. A riscv64 port would need either an InfluxData engineering decision to extend the CircleCI matrix, or an external contributor to submit and shepherd a PR through InfluxData’s CLA-gated, maintainer-controlled review process – neither has happened to date.

Acceptance probability: given the permissive MIT/Apache-2.0 license, the absence of any stated objection, and the fact that most of the Rust dependency stack is already riscv64-clean (Section 9), a well-formed riscv64 CI/build-matrix PR has no apparent technical reason to be rejected. However, with zero prior riscv64 activity, zero RISE involvement, and a purely reactive InfluxData review process, there is no evidence of institutional appetite to prioritize this work absent an external submission.

13. Readiness Assessment

  • Color: orange (downstream-only)
  • Release provider: distro (Ubuntu) – and this release is a stale, third-party legacy artifact, not a current upstream build
  • Justification: Upstream influxdata/influxdb has no riscv64 CI (.circleci/config.yml, direct-read confirmed zero riscv64 references) and publishes no riscv64 release binaries in any of its last five releases (influxdata/influxdb releases). Per the distribution floor rule, the presence of an Ubuntu-packaged influxdb riscv64 binary (packages.ubuntu.com/resolute/riscv64/influxdb) would ordinarily lift the project to yellow, but that package is an eight-year-old, unmaintained InfluxDB 1.6 release-candidate snapshot with no confirmed patch status and no evidence it has been verified to run correctly – so the “patch status unknown” condition applies, capping the grade at orange (downstream-only) rather than yellow.
  • InfluxDB is not an optimization-purpose project (it is a database server, not a SIMD/allocator/compression library whose sole value proposition is raw computational speed), so the Step 2 optimization-level cap does not apply; Optimization level is omitted from the header per the skill’s own scoping rule.
  • Pending work that could change the grade: none identified. There is no open riscv64 PR, no RISE engagement (InfluxData/InfluxDB is not a RISE member and has no RISE blog coverage, wheel-builder entry, or runner usage), and the only trace of forward motion is an unactioned entry in this repository’s own internal project-evaluation backlog (project-reports/.queue.yml) queuing InfluxDB as a future candidate for evaluation – not evidence of any external or upstream activity.

14. Investment Analysis

RISE has not funded or performed any work on InfluxDB riscv64 enablement to date (Section 12/13); nothing below is already covered.

14.1 Functional Enablement

  • Add a riscv64gc-unknown-linux-gnu target to the CircleCI build matrix in .circleci/config.yml and validate a cargo build --profile release completes cleanly.
  • Feature-gate or disable the WASM UDF-hosting feature (wasmtime/cranelift-codegen) on riscv64 until upstream Cranelift riscv64 codegen bugs (#13959, #13078, #12195, #9186) are resolved, or track and validate upstream fixes as they land.
  • Validate the full Rust dependency stack (Arrow, DataFusion, PyO3, sqlx/SQLite, jemalloc, BLAKE3, zstd, brotli) actually builds and passes tests end-to-end on riscv64 hardware – individual crates look clean per Section 9, but no one has run the full InfluxDB 3 test suite on riscv64.
  • Close or update the stale #20559 boltdb/bbolt issue to remove the misleading “riscv64 blocked by boltdb” narrative that downstream packagers (Alpine) are still propagating.

14.2 Performance Optimization

  • Not a current priority: functional enablement has not happened, so there is nothing to optimize yet. Once functional, BLAKE3’s lack of a riscv64 vector SIMD codepath (BLAKE3-team/BLAKE3 #484) is the one identified, quantifiable-in-principle performance gap (reported 3-5x slower on xof benchmarks vs an accelerated architecture), though no InfluxDB-level benchmark exists to translate that into an end-to-end database performance delta.

14.3 CI/CD Infrastructure

  • Add a riscv64 executor/target to CircleCI (.circleci/config.yml, core/.circleci/config.yml) mirroring the existing arm64 build+test+package-validation jobs.
  • Add riscv64 to .circleci/packages/config.yaml for .deb/.rpm package validation, and to .circleci/scripts/fetch-python-standalone.bash’s embedded-Python target-fetch list.
  • Determine whether native riscv64 CI hardware (e.g., RISE-provided runners) or QEMU emulation is used, given real-hardware-only reproduction of some Cranelift bugs (Section 9).

14.4 Ecosystem Enablement

Not applicable – Section 10 is omitted per the report template’s rule, since InfluxDB (the server) has no significant dependent package ecosystem requiring separate riscv64 enablement; its Python client (influxdb on PyPI) is a pure-Python, architecture-independent package that already works on riscv64 by construction.

14.5 Summary Table

Area Work Item Effort (person-weeks) Owner Priority
Functional Add riscv64gc-unknown-linux-gnu target to CircleCI build matrix, validate clean cargo build --release 1-2 InfluxData or external contributor Critical
Functional Validate full Rust dependency stack builds/tests on riscv64 hardware end to end 2-3 InfluxData or external contributor Critical
Functional Feature-gate WASM UDF hosting (wasmtime/Cranelift) off on riscv64 pending upstream fixes 1 InfluxData High
Functional Track/validate upstream wasmtime riscv64 Cranelift bug fixes as they land, re-enable WASM UDF feature 1-2 (ongoing) InfluxData or external contributor Medium
CI/CD Add riscv64 executor and package-validation (deb/rpm) jobs to CircleCI, mirroring arm64 1-2 InfluxData High
CI/CD Add riscv64 official release assets (linux binary, .deb, .rpm) once CI is green 0.5-1 InfluxData High
Ecosystem Correct/close stale boltdb-blocker narrative in #20559 to stop downstream packagers propagating it 0.1 InfluxData Low

15. Updates

(No updates yet – initial report dated 2026-09-07.)

16. References