Field notes from the IoT data plane.
Product news, engineering decisions, and practical guidance for teams running device infrastructure.
Before a standby can take traffic
Exclude the old writer, acquire the current term, replay durable state and require fresh whole-node readiness.
Read article ↗Read the workload behind a broker measurement
Use exact images, offered traffic, sink receipts and thermal conditions to judge whether a measurement applies to your deployment.
Read article ↗Follow an operation from request to evidence
Use the panel or your own agent to inspect current scope and follow a single command through daemon-owned reconciliation.
Read article ↗Keep ingress and recovery ownership clear
Reuse nginx connection and TLS operations while keeping shared message state and external delivery behind explicit process boundaries.
Read article ↗Choose around the state and operations you own
Map live readings, restart recovery, integrations and standby operation to the product scope your application needs.
Read article ↗Plan a Mosquitto or EMQX cutover
Map effective behavior, test an isolated candidate and preserve state and rollback decisions before assigning traffic.
Read article ↗