Hot restart¶

Ease of operation is one of the primary goals of Envoy. In addition to robust statistics and a local administration interface, Envoy has the ability to “hot” or “live” restart itself. This means that Envoy can fully reload itself (both code and configuration) without dropping any connections. The hot restart functionality has the following general architecture:

• Statistics and some locks are kept in a shared memory region. This means that gauges will be consistent across both processes as restart is taking place.
• The two active processes communicate with each other over unix domain sockets using a basic RPC protocol.
• The new process fully initializes itself (loads the configuration, does an initial service discovery and health checking phase, etc.) before it asks for copies of the listen sockets from the old process. The new process starts listening and then tells the old process to start draining.
• During the draining phase, the old process attempts to gracefully close existing connections. How this is done depends on the configured filters. The drain time is configurable via the --drain-time-s option and as more time passes draining becomes more aggressive.
• After drain sequence, the new Envoy process tells the old Envoy process to shut itself down. This time is configurable via the --parent-shutdown-time-s option.
• Envoy’s hot restart support was designed so that it will work correctly even if the new Envoy process and the old Envoy process are running inside different containers. Communication between the processes takes place only using unix domain sockets.
• An example restarter/parent process written in Python is included in the source distribution. This parent process is usable with standard process control utilities such as monit/runit/etc.