Diego v1.31.1

Changes from v1.31.0 to v1.31.1

• Verified with garden-runc-release v1.10.0.
• Verified with garden-windows-bosh-release v0.13.0.
• Verified with etcd-release v117.
• Verified with cf-mysql-release v35.
• Verified with cflinuxfs2-release v1.176.0.

NOTE: This version of Diego fixes the issue with miscalculating container memory metrics and the resulting problems with cf push reported on the updated Diego v1.31.0 release notes and on the cf-dev mailing list. If you have applied the workaround documented there, you can remove it with this release.

Significant changes

Container Execution

• As a CF app developer, I expect to see the garden error when my app or task container failed to be created so that I can correct or assist in diagnosing the error
• As a CF app developer, I expect to see informative error messages in logs and events when downloads or uploads fail so that I can report the failure mode

cfdot

• As a Diego operator, I expect cfdot to accept generic TLS client credential flags for BBS and locket communication so that I can simplify the required TLS configuration

Routing

• Operator can discover metrics on the latency of gorouter routing table convergence

Per-Instance Proxy (Experimental)

• As a CF operator, I expect the instance Envoy proxy listeners to support only suitably strong, RSA-compatible cipher suites so that they are compatible with the RSA-based instance-identity credentials

App Logging and Metrics

• As an app developer, I expect the cell reps not to mis-report the instance memory quota and usage when Envoy proxies are not enabled

Component Logging and Metrics

• As a CF operator, I expect each route-emitter to emit its count of NATS messages sent to the gorouters so that I can determine any discrepancies with the message rate the gorouters receive
• As a Diego operator, I expect to see metrics about the number of Tasks started, succeeded, and failed on each cell VM so that I can understand whether certain cells are failing a significant proportion of work
• As a Diego operator, I expect to see request latency metrics for the locket server (in flight)
• As a Diego contributor, I expect scripts/check-metrics-documentation to work with the new loggregator client so that I can be sure component metrics are documented adequately
• As a Diego operator, I expect the count of failed Locket DB queries to be constant under normal operating conditions so that I do not think the system is behaving incorrectly

Test Suites and Tooling

• Allocate ports to every server that we spin up in unit/integration tests in a well-defined manner and location

Documentation

• cloudfoundry/diego-release #352: Tls notes
• cloudfoundry/diego-release #369: bosh-lite instructions confusing

BOSH job changes

None.

BOSH property changes

cfdot

• Deprecated diego.cfdot.bbs.ca_cert in favor of tls.ca_certificate.
• Deprecated diego.cfdot.bbs.client_cert in favor of tls.certificate.
• Deprecated diego.cfdot.bbs.client_key in favor of tls.private_key.

rep and rep_windows

• Added admin_api.require_tls: whether to require TLS for the rep's admin API server.

BOSH link changes

None.