v1.23.0

🔗 Feature

• Adds an address field for the metrics server to allow external scraping #62 Thanks @tlwr

🔗 Updates

• Core-dns changed some metric names
• Updated dependencies

v1.22.0

🔗 feature

• Add an optional upcheck validation for internal domains, thanks @beyhan #59

🔗 dependency bumps

• Bump dependencies (Golang, vendored libraries)

v1.21.0

Feature

• coredns metrics integration (thanks @beyhan and @s4heid)
  • enable with metrics.enabled (default false)
  • listens on metrics.port (default 53088) on 127.0.0.1

Updates

• Bump dependencies (Golang, vendored libraries)

v1.20.0

Fixes

• bosh-dns may stop resolving internal domains until process manually restarted (#55)
• post-start script waits for bosh-dns and bosh-dns-resolvconf to be ready
• race condition on startup where queries for internal domains may be forwarded to recursor

Updates

• Bump dependencies (Golang, vendored libraries)

v1.19.0

Known issue

• bosh-dns may stop resolving internal domains until process manually restarted (#55) (fixed in v1.20.0)

Features/Updates

• Bump dependencies (vendored libraries)

Fixes

• bosh-dns fails to bind to 127.0.0.1 on windows: #171637356

v1.18.0

Known Issues

• bosh-dns fails to bind to 127.0.0.1 on windows: #171637356 (fixed in v1.19.0)
• bosh-dns may stop resolving internal domains until process manually restarted (#55) (fixed in v1.20.0)

Features/Updates

• Implement reverse DNS lookups for deployed VMs #170963966
• Skip smart health checks for single records #169610417
• Add port and log_level to windows spec
• Add addresses_files_glob to linux spec
• All DNS requests/responses logged at debug
• Bump dependencies (Golang, vendored libraries)

Fixes

• Avoid SERVFAIL responses: #170937906
  • servfail causes clients to try the next recursor and should be avoided
  • return nxdomain if dns blob has no matching domain
  • return formerr if query/domain can't be parsed
  • return success with 0 answers if no domains match query filter
• 5 second timeout for graceful dns server shutdown #171005042
• Try next recursor for non-successful responses #171082087
• Synchronous health checks should wait for inflight health requests #171163168
• Fix answer shuffling for small result sets #171185338

v1.17.0

Features/Updates

• Bump dependencies (Golang, Ruby, caddy, vendored libraries)
• log dns queries and health check responses at debug level, failures at warn level
• new configurable fields:
  • local_health_interval: how often the health server polls monit and local job health scripts
  • remote_health_interval: how often bosh-dns polls remote health servers
  • synchronous_check_timeout: network timeout for queries with y1 when health status not cached
  • request_timeout: network timeout for health server requests

Fixes

• stop redundant health check retries between poll intervals
• stop redundant health check requests when querying a domain
• Fix compilation error on windows due to broken symlinks
• health server should shutdown cleanly on sigterm

v1.16.0

Features/Updates

• Bump dependencies (Golang, Ruby, vendored libraries)
• Stop sending duplicate health check requests if a first request is still in flight
• Apply consistent handling for EDNS, truncation and compression of DNS responses
• Force TLS 1.2 to mitigate performance regression seen with TLS 1.3

v1.15.0

Features/Updates

• Bump dependencies (Golang, Ruby, vendored libraries)
• Resolve BOSH_AGENT_ID.bosh-agent-id queries
• Make log_level configurable (defaults to DEBUG)

v1.12.0

Features/Updates

• Deployment operator can specify an option to change the recursor selection to be in order of ns servers specified in resolv.conf #165801868
• Make use of SO_REUSEPORT #165299945 [linux only]