BoltPipeline logo
Agent Docs|All Docs

Environments

The agent runs in one of four environments: dev, lab, int, or prd. Each has a distinct execution path and feature set.

Overview

Four environments, two execution paths

BoltPipeline uses two execution models. The environment you configure determines which path the agent follows — and whether DDL Deploy is available.

PECOint and prd

Plan, Enrich, Certify, Operate

The production path. Schema changes travel through the promotion workflow — not deployed directly by the agent.

SQL must be certified (read-only validation) before it can be operated.

DDL Deploy is disabled — schema must already match when certify runs.

PECO/Ddev and lab

Plan, Enrich, Certify, Deploy, re-Certify, Operate

The development path. If Certify detects schema gaps (missing tables or columns), DDL Deploy fixes them automatically.

After DDL Deploy, Certify re-runs to confirm the schema is now correct, then Operate proceeds.

Enables fast iteration — write SQL, certify, deploy schema, re-certify, operate — in one cycle.

Reference

Environment details

devPECO/DPrimary development tier
Execution pathPECO/D — DDL Deploy enabled
DDL DeployYes
File credentialsAllowed
Log format defaulttext
labPECO/DSandbox / demo tier
Execution pathPECO/D — DDL Deploy enabled
DDL DeployYes
File credentialsAllowed
Log format defaulttext
intPECOIntegration / pre-production tier
Execution pathPECO — no DDL Deploy
DDL DeployNo — schema via promotion
File credentialsAllowed
Log format defaultjson recommended
prdPECOProduction tier
Execution pathPECO — no DDL Deploy
DDL DeployNo
File credentialsBLOCKED — secret manager required
Log formatjson (enforced)

Reference

Feature matrix

Featuredevlabintprd
PECO/D (DDL Deploy)✓✓✗✗
PECO (certified Operate)✓✓✓✓
File-based credentials✓✓✓blocked
Secret manager credentials✓✓✓✓
Text log format✓✓✓✗
JSON log format✓✓✓✓ (enforced)
Discovery + Profile✓✓✓✓
Certifier (read-only)✓✓✓✓
Offline spool✓✓✓✓

Guidance

Which environment to use

Use dev when:
  • •Writing new pipelines or modifying existing SQL
  • •You need DDL Deploy to fix schema gaps automatically
  • •First-time setup and onboarding
  • •Iterating quickly on data transformations
Use lab when:
  • •Demo and sandbox work on isolated data
  • •Parallel iteration without touching dev data
  • •Feature development that shouldn't affect dev pipelines
Use int when:
  • •Pre-production validation of promoted pipelines
  • •Integration testing against integration-tier data
  • •Verifying certified SQL runs correctly before prd
  • •Smoke testing after a promotion
Use prd when:
  • •Running certified, promoted pipelines in production
  • •Requires secret manager — file-based credentials are blocked
  • •JSON logging and structured observability
  • •Never use for development or schema iteration

Each agent instance is tied to one environment. The bootstrap package you download from the Console is environment-specific — the agent will only execute pipelines for that environment. To run multiple environments, run multiple agent containers (one per environment), each with its own config file and bootstrap package.

Need help?

Email us at support@boltpipeline.ai