agentleFS
NVIDIAKnown publisherNVIDIA/nvcf/src/control-plane-services/function-autoscaler/AGENTS.md

nvcf / function-autoscaler

AGENTS.md261 starsChanged 9 days agoApache-2.0View on GitHub

The cargo run above has the same Bazel-only dependency described under Build & Test, so it does not resolve either. Run the :server Bazel target instead. crates/server depends on //src/libraries/rust/nvcf-info as a Bazel target only, and cargo does not declare it, so cargo build, cargo clippy, and cargo test fail to resolve it. Build and test with Bazel from the monorepo root: Formatting and advisory checks do not compile the crate, so they still run under cargo: CI subproject id:…

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    All operational parameters are externalized to YAML settings files with #[serde(default)].

    line 27

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

AGENTS.md

Repo Role

  • Repo: rs-autoscaler
  • Language: Rust
  • Purpose: Autoscaling service for NVCF monitors function utilization via a timeseries database and adjusts instance counts through the NVCF API.

Architecture

  • Scaling loop: Queries utilization metrics, applies per-function scaling policies (static or custom via gRPC), decides desired instance count.
  • Discovery: Finds active functions from the timeseries DB (invocations + workers), writes them to Cassandra.
  • Policy cache: Fetches per-function scaling configs from a gRPC service, caches with configurable TTL.
  • Distributed locking: Cassandra-based locks for coordinating discovery and scaling across replicas.

Key Paths

  • crates/server/src/scaling/: scaling logic, policy client/cache, thresholds, stickiness
  • crates/server/src/work/: scaling loop, discovery, utilization queries
  • crates/server/src/cassandra/: Cassandra client, statements, distributed locks
  • crates/server/src/timeseries_db/: timeseries DB client (Thanos/VictoriaMetrics compatible)
  • crates/server/src/nvcf_api/: NVCF API client, OAuth2 token client
  • crates/server/resources/: YAML config files per environment

Config

  • All operational parameters are externalized to YAML settings files with #[serde(default)].
  • Environment variable overrides use __ as separator (e.g., SCALING__UTILIZATION_WINDOW_SECONDS=70).
  • Secrets are loaded from a JSON file at the path specified by secrets_path in config.

Local Development

# Start dependencies
docker compose -f crates/server/nvcf-service/local_env/docker-compose.yml up -d cassandra victoriametrics

# Copy and fill secrets
cp local_env/secrets/secrets.json.tmpl local_env/vault/secrets.json

# Run with local config
CONFIG=crates/server/resources/settings-local.yaml \
SECRETS_PATH=local_env/vault/secrets.json \
cargo run --bin server -p rs-autoscaler

The cargo run above has the same Bazel-only dependency described under Build & Test, so it does not resolve either. Run the :server Bazel target instead.

Build & Test

crates/server depends on //src/libraries/rust/nvcf-info as a Bazel target only, and cargo does not declare it, so cargo build, cargo clippy, and cargo test fail to resolve it. Build and test with Bazel from the monorepo root:

bazel test //src/control-plane-services/function-autoscaler/crates/server:all

Formatting and advisory checks do not compile the crate, so they still run under cargo:

cargo fmt -p rs-autoscaler
cargo deny check advisories

CI subproject id: function-autoscaler. Follow the root RELEASE.md and docs/dev/github-release-process.md for source versioning and publication handoffs. Inspect .github/workflows/ for current public validation.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.