Solarix

A universal Solana indexer in Rust — point it at any Anchor program and query its on-chain data over REST.

Give Solarix any Anchor program ID and it fetches the IDL directly from the chain, generates a typed PostgreSQL schema at runtime — no codegen, no recompile, no redeploy — then indexes transactions and account state through a concurrent backfill-plus-streaming pipeline. Decoded data is immediately queryable through a 13-endpoint REST API with typed filters, cursor pagination, and time-series aggregations. Built for the Superteam Ukraine bounty (Middle level, 500 USDG).

What's notable here:

Runtime schema generation — PostgreSQL DDL built from the IDL at registration; any Anchor program works on the first request
Hybrid typed + JSONB storage — scalars promoted to native columns for fast filters, complex types preserved without loss in JSONB
Concurrent backfill + streaming — historical RPC and live WebSocket run at once; idempotent writes make the overlap harmless
Typed errors, zero unwrap — retry classification the compiler enforces, behind 383 tests and strict clippy

POST /api/programs { "program_id": "LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo" }
Solarix: fetch IDL on-chain --> CREATE SCHEMA + TABLES --> backfill + stream --> query API ready

Architecture

flowchart LR
RPC["Solana RPC<br/>HTTP JSON-RPC"]
WS["Solana WS<br/>logsSubscribe"]
subgraph READ["Read Layer"]
R1[RpcClient]
R2[WsTransactionStream]
end
subgraph DECODE["Decode Layer"]
D1["ChainparserDecoder<br/>Borsh to JSON"]
end
subgraph STORE["Store Layer"]
S1["StorageWriter<br/>batch INSERT + upsert"]
end
subgraph SERVE["Serve Layer"]
A1["axum Router<br/>13 endpoints"]
end
PG[("PostgreSQL 16<br/>typed + JSONB")]
CLI["REST Clients"]
RPC -->|getBlock| R1
WS -->|logsSubscribe| R2
R1 -->|"mpsc (256)"| D1
R2 -->|"mpsc (256)"| D1
D1 -->|"mpsc (256)"| S1
S1 --> PG
PG --> A1
A1 --> CLI

The pipeline has four layers connected by bounded Tokio channels (capacity 256). Solana data enters through HTTP RPC (for historical blocks) and WebSocket (for live transactions), flows through the Borsh decoder, gets written to PostgreSQL, and is served by the axum REST API.

stateDiagram-v2
[*] --> Initializing
Initializing --> Backfilling : checkpoint behind chain tip
Initializing --> Streaming : no gap detected
Backfilling --> Streaming : caught up to tip
Streaming --> CatchingUp : gap detected
CatchingUp --> Streaming : gap filled
Streaming --> ShuttingDown : SIGTERM or SIGINT
ShuttingDown --> [*] : drain channels, flush DB

The pipeline starts in Initializing, transitions to Backfilling if there is a historical gap, then enters Streaming. A WebSocket slot gap triggers CatchingUp for a mini-backfill. SIGTERM begins a graceful shutdown: in-flight messages drain, final writes flush to PostgreSQL.

Cold Start Strategy

On cold start, Solarix runs backfill and streaming concurrently (Option C). Both write to the same tables with INSERT ON CONFLICT DO NOTHING, so duplicate processing is harmless and crash recovery is automatic. Live data is available immediately while historical backfill catches up in the background.

Database Layout

Each registered program gets its own PostgreSQL schema:

public/
programs -- program registry (program_id, schema_name, idl_hash, status)
indexer_state -- per-program pipeline state (last_slot, status, counters)
{program_name}_{program_id_prefix}/ -- e.g. lb_clmm_lbuzkhrx/
lbpair -- one table per account type, upsert on pubkey
position -- (promoted typed columns + JSONB data column)
binarray
...
_instructions -- all decoded instructions (append-only)
_checkpoints -- slot cursor per stream (backfill, stream)
_metadata -- field names and types from IDL

Simple scalars (u64, bool, String, Pubkey) are promoted to native PostgreSQL columns for fast indexed queries. Complex nested types live in a JSONB data column with GIN indexes. Every field is preserved regardless of promotion — there is no data loss on schema generation.

Source module tree (20 modules, 14,109 lines)

src/
main.rs Entry point: signal handling, pipeline + API startup
lib.rs Public module declarations
config.rs 22 env vars via clap, validation
types.rs DecodedInstruction, DecodedAccount, BlockData, TransactionData
registry.rs Two-phase program registration state machine
runtime_stats.rs Process-wide atomic counters for Prometheus metrics
startup.rs Router + AppState construction, extracted for testability
idl/
mod.rs IdlManager: cache, parse, validate (v0.30+ only)
fetch.rs Fetch cascade: on-chain PDA -> bundled -> manual upload
decoder/
mod.rs ChainparserDecoder: Borsh deserializer for 18+ IDL types
pipeline/
mod.rs PipelineOrchestrator: 5-state machine, concurrent backfill+stream
rpc.rs RPC client with rate limiting (governor) and retry (backon)
ws.rs WebSocket logsSubscribe with dedup cache and heartbeat
storage/
mod.rs DB pool init, system table bootstrap
schema.rs IDL -> CREATE TABLE/INDEX DDL, promoted column detection
writer.rs Batch INSERT...UNNEST, account upsert, checkpoint management
queries.rs Dynamic query builder for API filters
api/
mod.rs axum Router, AppState, ApiError -> HTTP status mapping
handlers.rs 13 endpoint handlers with pagination and cursor encoding
filters.rs Filter parsing, operator validation against IDL

For the full architecture deep-dive, see docs/architecture.md.

Technology Stack

Crate	Version	Purpose
`axum`	0.8	HTTP framework and router
`sqlx`	0.8	Async PostgreSQL driver
`tokio`	1 (latest)	Async runtime
`anchor-lang-idl-spec`	0.1	Anchor IDL type definitions
`tokio-tungstenite`	0.26	WebSocket client for `logsSubscribe`
`governor`	0.10	Async GCRA rate limiting
`backon`	1	Exponential backoff retry
`thiserror`	2	Typed error enums
`tracing` + `tracing-subscriber`	0.1	Structured logging (JSON + pretty)
`metrics` + `metrics-exporter-prometheus`	0.24 / 0.17	Prometheus metrics endpoint
`clap`	4	CLI and environment variable configuration
`sha2`	0.10	Instruction/account discriminator computation
`flate2`	1	IDL zlib decompression
`tower-http`	0.6	Request ID propagation and tracing middleware

Quick Start

Docker Compose (Recommended)

Clone the repository and enter the directory:

git clone https://github.com/valentynkit/solarix.git
cd solarix

Start the full stack (PostgreSQL 16 + Solarix):
```
docker compose up --build
```
Verify the API is ready:
```
curl -s http://localhost:3000/health | jq
```
Register a program — Solarix fetches the IDL on-chain, creates the schema, and starts indexing. The example below uses Meteora DLMM because it has rich on-chain activity and was verified end-to-end during the Sprint-4 e2e gate (199 swaps indexed in 5 minutes):
```
curl -s -X POST http://localhost:3000/api/programs \
-H "Content-Type: application/json" \
-d '{"program_id":"LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo"}'| jq
```
```
{
"data": {
"program_id": "LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo",
"status": "schema_created",
"idl_source": "onchain"
},
"meta": {
"message": "Program registered. Indexing will begin on next restart."
}
}
```
The schema is generated synchronously during the request. GET /api/programs/{id} will return the same "status": "schema_created" immediately after, along with "schema_name": "lb_clmm_lbuzkhrx".
Restart the service so the pipeline auto-starts for the registered program:
```
docker compose restart solarix
```

Query decoded data once indexing begins (allow a minute for the first slot to appear):

# Decoded swap instructions (cursor-paginated)
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap?limit=5"| jq
# Filter: swaps where amount_in > 0.001 SOL (1,000,000 lamports)
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap?data.amount_in_gt=1000000&limit=10"| jq
# Time-series swap count by hour
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap/count?interval=hour"| jq
# Indexing statistics
curl -s http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/stats | jq

Run the automated end-to-end demo:
```
bash demo.sh
```

Local Development (No Docker)

Start a local PostgreSQL instance:

docker run -d --name solarix-db \
-e POSTGRES_DB=solarix \
-e POSTGRES_USER=solarix \
-e POSTGRES_PASSWORD=solarix \
-p 5432:5432 \
postgres:16

Configure environment variables and run:

export DATABASE_URL="postgres://solarix:solarix@localhost:5432/solarix"export SOLANA_RPC_URL="https://api.devnet.solana.com"
cargo run

Build, test, and lint:

cargo build # debug build
cargo build --release # optimized build
cargo watch -x run # hot-reload during development
cargo test# 383 tests across all suites (MSRV 1.88)
cargo test --features integration # integration tests (requires Docker)
cargo clippy # strict lints (unwrap/expect/panic denied)
cargo fmt -- --check # formatting check

API Reference

Base URL: http://localhost:3000

Programs

Method	Endpoint	Description
`POST`	`/api/programs`	Register a program (auto-fetches IDL or accepts manual upload)
`GET`	`/api/programs`	List all registered programs
`GET`	`/api/programs/{id}`	Get program details and schema info
`DELETE`	`/api/programs/{id}?drop_tables=true`	Deregister a program (optionally drop its schema)
`GET`	`/api/programs/{id}/stats`	Indexing statistics (total instructions, accounts, last slot)

# Register a program
curl -s -X POST http://localhost:3000/api/programs \
-H "Content-Type: application/json" \
-d '{"program_id":"LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo"}'| jq
# {# "data": {# "program_id": "LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo",# "status": "schema_created",# "idl_source": "onchain"# },# "meta": { "message": "Program registered. Indexing will begin on next restart." }# }# Indexing statistics
curl -s http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/stats | jq
# {# "data": {# "program_id": "LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo",# "total_instructions": 142831,# "total_accounts": 9204,# "last_processed_slot": 318472910# }# }

Instructions & Accounts

Method	Endpoint	Description
`GET`	`/api/programs/{id}/instructions`	List instruction types from IDL
`GET`	`/api/programs/{id}/instructions/{name}`	Query decoded instructions with filters
`GET`	`/api/programs/{id}/instructions/{name}/count`	Count instructions, optionally by time interval
`GET`	`/api/programs/{id}/accounts`	List account types from IDL
`GET`	`/api/programs/{id}/accounts/{type}`	Query decoded accounts with filters
`GET`	`/api/programs/{id}/accounts/{type}/{pubkey}`	Get a single account by pubkey

# Query decoded swap instructions (cursor pagination)
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap?limit=5"| jq
# Filter: swaps with amount_in > 0.001 SOL
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap?data.amount_in_gt=1000000&limit=10"| jq
# Time-series count by hour
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap/count?interval=hour"| jq
# List account types
curl -s http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/accounts | jq
# Query accounts with offset pagination
curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/accounts/lbpair?limit=5&offset=0"| jq

Observability

Method	Endpoint	Description
`GET`	`/health`	System health with DB connectivity and per-program pipeline status
`GET`	`/metrics`	Prometheus metrics (requires `SOLARIX_METRICS_ENABLED=true`)

# Health check
curl -s http://localhost:3000/health | jq
# {# "status": "healthy",# "database": "connected",# "programs": [# {# "program_id": "LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo",# "status": "streaming",# "last_processed_slot": 318472910# }# ]# }# Prometheus metrics (enable with SOLARIX_METRICS_ENABLED=true)
curl -s http://localhost:3000/metrics | grep "^solarix_"

Filter Syntax

Filters are appended as query parameters directly to any instruction or account query endpoint:

?data.amount_gt=1000000 # JSONB field, greater than?data.authority_eq=So11111111111111111111111111111111111111112 # equals (full address)?data.is_active_eq=true # boolean?slot_gt=300000000 # promoted BIGINT column (no "data." prefix)

Combine multiple filters with &:

?data.amount_in_gt=1000000&data.min_amount_out_lt=5000000000

Operator	Meaning
`_eq`	Equals
`_ne`	Not equals
`_gt`	Greater than
`_gte`	Greater than or equal
`_lt`	Less than
`_lte`	Less than or equal
`_in`	In list (comma-separated)
`_contains`	JSONB containment (`@>`) — matches rows whose JSON value contains the argument

Pagination

?limit=50&offset=0 # Offset-based (accounts)?limit=50&cursor=abc123 # Cursor-based (instructions, stable on append-only data)

Error Responses

All errors return structured JSON:

curl -s "http://localhost:3000/api/programs/LBUZKhRxPF3XUpBCjp4YzTKgLccjZhTSDM9YuVaPwxo/instructions/swap?nonexistent_gt=1"| jq

{
"error": {
"code": "INVALID_FILTER",
"message": "Unknown field 'nonexistent'. Available: amount_in, min_amount_out, ...",
"available_fields": ["amount_in", "min_amount_out"]
}
}

Code	HTTP Status	Meaning
`PROGRAM_NOT_FOUND`	404	Program ID not registered
`PROGRAM_ALREADY_REGISTERED`	409	Duplicate registration attempt
`INVALID_FILTER`	400	Bad filter syntax or unknown field (returns `available_fields`)
`INVALID_REQUEST`	400	Malformed request body
`IDL_ERROR`	422	IDL fetch/parse failure
`STORAGE_ERROR`	500	Database error
`QUERY_FAILED`	500	Query execution error

Configuration

All parameters are configured via environment variables (or CLI flags). Copy .env.example to .env and customize. Only DATABASE_URL is required.

Core

Variable	Type	Default	Description
`DATABASE_URL`	String	(required)	PostgreSQL connection string
`SOLANA_RPC_URL`	String	`https://api.mainnet-beta.solana.com`	Solana JSON-RPC endpoint
`SOLANA_WS_URL`	String	(derived from RPC URL)	WebSocket endpoint for `logsSubscribe`

Database

Variable	Type	Default	Description
`SOLARIX_DB_POOL_MIN`	u64	`2`	Minimum connection pool size
`SOLARIX_DB_POOL_MAX`	u64	`10`	Maximum connection pool size

API

Variable	Type	Default	Description
`SOLARIX_API_HOST`	String	`0.0.0.0`	Bind address
`SOLARIX_API_PORT`	u64	`3000`	Bind port
`SOLARIX_API_PAGE_SIZE`	u64	`50`	Default page size
`SOLARIX_API_MAX_PAGE_SIZE`	u64	`1000`	Maximum allowed page size

Pipeline

Variable	Type	Default	Description
`SOLARIX_RPC_RPS`	u64	`10`	RPC rate limit (requests/second)
`SOLARIX_BACKFILL_CHUNK_SIZE`	u64	`50000`	Slots per backfill batch
`SOLARIX_START_SLOT`	u64	(auto)	Override backfill start slot
`SOLARIX_END_SLOT`	u64	(auto)	Override backfill end slot
`SOLARIX_INDEX_FAILED_TXS`	bool	`false`	Index failed transactions
`SOLARIX_CHANNEL_CAPACITY`	u64	`256`	Bounded channel size between pipeline stages
`SOLARIX_CHECKPOINT_INTERVAL_SECS`	seconds	`10`	Checkpoint persistence interval

Retry & Resilience

Variable	Type	Default	Description
`SOLARIX_RETRY_INITIAL_MS`	u64	`500`	Initial retry backoff (ms)
`SOLARIX_RETRY_MAX_MS`	u64	`30000`	Maximum retry backoff (ms)
`SOLARIX_RETRY_TIMEOUT_SECS`	seconds	`300`	Total retry timeout before giving up
`SOLARIX_MAX_CONSECUTIVE_FETCH_FAILURES`	u64	`100`	Max consecutive RPC failures before halt
`SOLARIX_SHUTDOWN_DRAIN_SECS`	seconds	`15`	In-flight message drain timeout on shutdown
`SOLARIX_SHUTDOWN_DB_FLUSH_SECS`	seconds	`10`	Final DB write timeout on shutdown

WebSocket

Variable	Type	Default	Description
`SOLARIX_WS_PING_INTERVAL_SECS`	seconds	`30`	Heartbeat ping interval
`SOLARIX_WS_PONG_TIMEOUT_SECS`	seconds	`10`	Pong response timeout
`SOLARIX_DEDUP_CACHE_SIZE`	u64	`10000`	Signature dedup cache capacity

Logging

Variable	Type	Default	Description
`SOLARIX_LOG_LEVEL`	String	`info`	Log level (`trace`, `debug`, `info`, `warn`, `error`)
`SOLARIX_LOG_FORMAT`	String	`json`	Log format (`json` for production, `pretty` for development)

Observability

Variable	Type	Default	Description
`SOLARIX_METRICS_ENABLED`	bool	`false`	Enable Prometheus `/metrics` endpoint
`SOLARIX_METRICS_PATH`	String	`/metrics`	Path for the Prometheus metrics endpoint

Architectural Decisions

1. Runtime schema generation

PostgreSQL DDL is generated directly from the Anchor IDL at program registration time — no code generation step required.

Alternatives: compile-time codegen (anchor-gen, seahorse) requires a recompile and redeploy for every new program; predefined tables hardcode specific programs and break universality.

Why: Zero additional configuration. The IDL is the only input. Any Anchor program works on the first POST /api/programs.

2. Hybrid typed + JSONB storage

Simple scalar fields (u64, bool, Pubkey, String) are promoted to native PostgreSQL columns with proper types. Complex types (structs, vecs, enums) live in a JSONB data column with a GIN index.

Alternatives: all-JSONB is simple but makes range queries slow and loses type coercion; fully normalised relational storage creates schema explosion for deeply nested types.

Why: Native columns give fast indexed queries on the most-filtered fields. JSONB fallback preserves every field regardless of type complexity, with no data loss on schema generation.

3. Concurrent backfill + streaming (Option C)

On cold start, historical backfill (HTTP RPC) and live streaming (WebSocket) run simultaneously. Both write to the same tables with INSERT ON CONFLICT DO NOTHING.

Alternatives: sequential (backfill then stream) leaves a gap window and delays live data; streaming-only loses all historical data.

Why: Live data is available immediately. Idempotent writes make overlap harmless. Crash recovery is automatic — both paths resume from their checkpoints independently.

4. Typed thiserror error enums

Five module-level error enums (IdlError, DecodeError, StorageError, PipelineError, ApiError) with explicit is_retryable() classification and clippy::expect_used = "deny".

Alternatives:anyhow erases the type needed for retry classification; Box<dyn Error> requires string matching to distinguish retryable from fatal errors.

Why: The compiler enforces exhaustive error handling at every call site. Retry logic in the pipeline depends on typed classification. Zero unwrap in production paths.

5. GCRA rate limiting via governor

All outbound RPC calls pass through an async-native Generic Cell Rate Algorithm limiter set to 10 RPS by default.

Alternatives:sleep-based throttle blocks the Tokio executor and wastes wall time; no rate limit causes 429 storms on public endpoints that corrupt backfill progress.

Why:governor is fully async (no blocking), precise, and handles the ~10 RPS public Solana RPC limit with configurable burst tolerance.

Sprint-4 integration testing fixed three critical bugs

Real end-to-end testing against mainnet surfaced issues that 257 unit tests had not caught:

IDL PDA derivation used find_program_address — Anchor v0.30 uses create_with_seed from a program signer. Fixed in commit 99567f2. Regression test: test_idl_address_derivation_matches_anchor_v030.
programs.idl_json not persisted — the IDL was stored in memory but not written to the database, so the pipeline never auto-restarted after a container restart. Fixed in commit 797bf74. Regression test: test_idl_json_persisted_and_loaded_on_restart.
BIGINT filter cast — promoted-column SQL filters were bound as TEXT. PostgreSQL rejected bigint > text at runtime. Fixed in commit 243a0de. Regression test: test_slot_gt_filter_uses_bigint_cast.

All three have regression tests in tests/regression_e2e_sprint4.rs and run on every CI push.

Testing

Unit + Property Tests

Run the full unit and property-based test suite:

cargo test

The decoder is verified with proptest: it generates random values for all 18+ supported Anchor IDL types, Borsh-serialises them, decodes via ChainparserDecoder, and asserts the JSON output matches field-for-field. The full suite runs 383 tests across 21 suites.

Integration Tests (require Docker)

Integration tests spin up a PostgreSQL 16 container per test via Testcontainers:

cargo test --features integration

Covers: schema generation against a real database, filter SQL execution (including the Sprint-4 BIGINT vs TEXT regression), the full register → decode → query roundtrip, and a LiteSVM end-to-end path that deploys a real Anchor program into an in-process Solana VM.

Mainnet Smoke (nightly CI)

cargo test --release --features mainnet-smoke -- mainnet_smoke

Run automatically by .github/workflows/nightly.yml. Not required for local development.

CI

Every push to main runs nine parallel jobs: lint, unit, integration, coverage, fuzz-smoke, security audit, docker smoke, MSRV check, and a toolchain matrix (stable + beta). See .github/workflows/ci.yml.

Future Work

Planned

These features are in active development:

Multi-program concurrent orchestration — index multiple programs simultaneously with a shared rate limiter and per-program JoinSet lifecycle management
Pipeline auto-start on API registration — POST /api/programs immediately spawns the pipeline without requiring a service restart

Post-Submission Ideas

Geyser/gRPC plugin transport — replace HTTP RPC polling with a Geyser plugin connection for lower-latency block delivery
GraphQL query layer — expose the same typed data through a GraphQL interface for richer client queries
Schema evolution — ALTER TABLE on IDL update when new fields are added to an existing program
Grafana dashboard out-of-box — bundle a preconfigured Grafana dashboard JSON for the Prometheus metrics
Legacy Anchor IDL v0.29 support — extend the fetch cascade for programs that predate the metadata.spec field
Signature-list batch mode — accept a list of transaction signatures for targeted backfill without slot-range scanning
WebSocket backpressure metrics — expose queue depth and drop rate through the /metrics endpoint

License

MIT — see LICENSE.

Built for the Superteam Ukraine bounty. Read the build story.

Name		Name	Last commit message	Last commit date
.github/workflows		.github/workflows
_bmad-output		_bmad-output
docs		docs
fuzz		fuzz
idls		idls
scripts		scripts
src		src
tests		tests
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
.gitleaks.toml		.gitleaks.toml
CLAUDE.md		CLAUDE.md
CONTRIBUTING.md		CONTRIBUTING.md
Cargo.lock		Cargo.lock
Cargo.toml		Cargo.toml
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
clippy.toml		clippy.toml
demo.sh		demo.sh
deny.toml		deny.toml
docker-compose.yml		docker-compose.yml
rust-toolchain.toml		rust-toolchain.toml
rustfmt.toml		rustfmt.toml
skills-lock.json		skills-lock.json

Navigation Menu

Repository files navigation

Solarix

Table of Contents

Architecture

Cold Start Strategy

Database Layout

Technology Stack

Quick Start

Docker Compose (Recommended)

Local Development (No Docker)

API Reference

Programs

Instructions & Accounts

Observability

Filter Syntax

Pagination

Error Responses

Configuration

Core

Database

API

Pipeline

Retry & Resilience

WebSocket

Logging

Observability

Architectural Decisions

Testing

Future Work

Planned

Post-Submission Ideas

License

About

Releases

Packages

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

Solarix

Table of Contents

Architecture

Cold Start Strategy

Database Layout

Technology Stack

Quick Start

Docker Compose (Recommended)

Local Development (No Docker)

API Reference

Programs

Instructions & Accounts

Observability

Filter Syntax

Pagination

Error Responses

Configuration

Core

Database

API

Pipeline

Retry & Resilience

WebSocket

Logging

Observability

Architectural Decisions

Testing

Future Work

Planned

Post-Submission Ideas

License

About

Topics

Resources

Contributing

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages