Skip to main content
All PandaProbe SDK behavior can be configured via environment variables. These are read at client initialization time.
VariableTypeDefaultRequiredDescription
PANDAPROBE_API_KEYstring—YesYour PandaProbe API key. Required when the SDK is enabled.
PANDAPROBE_PROJECT_NAMEstring—YesThe project name to associate traces with. Required when the SDK is enabled.
PANDAPROBE_ENDPOINTstringhttps://api.pandaprobe.comNoThe PandaProbe backend URL. Change this for self-hosted deployments.
PANDAPROBE_ENVIRONMENTstringNoneNoEnvironment tag attached to all traces (for example, production, staging, development).
PANDAPROBE_RELEASEstringNoneNoRelease or version tag attached to all traces (for example, v1.2.3, a git SHA).
PANDAPROBE_ENABLEDbooleantrueNoSet to false to completely disable the SDK. All operations become no-ops. Accepted values: true, 1, yes (enabled) or any other value (disabled).
PANDAPROBE_BATCH_SIZEinteger10NoNumber of items per flush batch sent to the backend.
PANDAPROBE_FLUSH_INTERVALfloat5.0NoSeconds between automatic background flushes.
PANDAPROBE_MAX_QUEUE_SIZEinteger1000NoMaximum number of items in the send queue. When full, the oldest item is dropped with a warning log.
PANDAPROBE_DEBUGbooleanfalseNoEnable verbose debug logging. Sets the pandaprobe logger to DEBUG level. Useful for troubleshooting.

​
Setting environment variables

export PANDAPROBE_API_KEY="pk_..."
export PANDAPROBE_PROJECT_NAME="my-project"
export PANDAPROBE_ENDPOINT="https://your-instance.com"
export PANDAPROBE_ENVIRONMENT="production"
The SDK reads environment variables via os.environ.get() at initialization time. Changes to environment variables after the client is created have no effect. Use pandaprobe.init() to reconfigure.

​
Boolean parsing

Boolean environment variables (PANDAPROBE_ENABLED, PANDAPROBE_DEBUG) accept true, 1, or yes (case-insensitive) as truthy. All other values are treated as falsy.