Merge branch 'master' into version-gauge

Signed-off-by: Jose Nino <[email protected]>

.circleci/config.yml

@@ -39,6 +39,8 @@ jobs:
        - checkout
        - run: ci/do_circle_ci.sh bazel.coverage
        - run: ci/coverage_publish.sh
+       - store_artifacts:
+           path: /build/envoy/generated/coverage
    format:
      docker:
        - image: lyft/envoy-build:8839608c2f4b0d7c8311361ee5c8785e543d517a

bazel/README.md

@@ -273,7 +273,13 @@ test/run_envoy_bazel_coverage.sh
 The summary results are printed to the standard output and the full coverage
 report is available in `generated/coverage/coverage.html`.
 
-In Travis, the master branch and pull requests by members of the Lyft organization will have coverage reports automatically uploaded to S3 when the coverage target successfully executes. The latest coverage report for master is available [here](https://s3.amazonaws.com/lyft-envoy/coverage/report-master/coverage.html). If an uploaded coverage report is available for a branch's build, a link can be found at the bottom of the output log in Travis.
+Coverage for every PR is available in Circle in the "artifacts" tab of the coverage job. You will
+need to navigate down and open "coverage.html" but then you can navigate per normal. NOTE: We
+have seen some issues with seeing the artifacts tab. If you can't see it, log out of Circle, and
+then log back in and it should start working.
+
+The latest coverage report for master is available
+[here](https://s3.amazonaws.com/lyft-envoy/coverage/report-master/coverage.html).
 
 # Cleaning the build and test artifacts
 

bazel/envoy_build_system.bzl

@@ -63,8 +63,8 @@ def envoy_test_linkopts():
         "@bazel_tools//tools/osx:darwin": [],
 
         # TODO(mattklein123): It's not great that we universally link against the following libs.
-        # In particular, -latomic is not needed on all platforms. Make this more granular.
-        "//conditions:default": ["-pthread", "-latomic"],
+        # In particular, -latomic and -lrt are not needed on all platforms. Make this more granular.
+        "//conditions:default": ["-pthread", "-latomic", "-lrt"],
     })
 
 # References to Envoy external dependencies should be wrapped with this function.

docs/configuration/cluster_manager/cluster.rst

@@ -182,6 +182,8 @@ dns_resolvers
   configuration. For cluster types other than *strict_dns* and *logical_dns* this setting is
   ignored.
 
+.. _config_cluster_manager_cluster_outlier_detection_summary:
+
 :ref:`outlier_detection <config_cluster_manager_cluster_outlier_detection>`
   *(optional, object)* If specified, outlier detection will be enabled for this upstream cluster.
   See the :ref:`architecture overview <arch_overview_outlier_detection>` for more information on outlier

docs/configuration/http_conn_man/stats.rst

@@ -68,3 +68,19 @@ the following statistics:
    downstream_cx_total, Counter, Total connections
    downstream_cx_destroy_remote_active_rq, Counter, Total connections destroyed remotely with 1+ active requests
    downstream_rq_total, Counter, Total requests
+
+
+Per listener statistics
+-----------------------
+
+Additional per listener statistics are rooted at *listener.<address>.http.<stat_prefix>.* with the
+following statistics:
+
+ .. csv-table::
+    :header: Name, Type, Description
+    :widths: 1, 1, 2
+
+    downstream_rq_2xx, Counter, Total 2xx responses
+    downstream_rq_3xx, Counter, Total 3xx responses
+    downstream_rq_4xx, Counter, Total 4xx responses
+    downstream_rq_5xx, Counter, Total 5xx responses

docs/configuration/http_filters/router_filter.rst

@@ -225,6 +225,17 @@ for the next health check interval. The host can become healthy again via standa
 checks. See the :ref:`health checking overview <arch_overview_health_checking>` for more
 information.
 
+.. _config_http_filters_router_x-envoy-overloaded:
+
+x-envoy-overloaded
+^^^^^^^^^^^^^^^^^^
+
+If this header is set by upstream, Envoy will not retry. Currently the value of the header is not
+looked at, only its presence. Additionally, Envoy will set this header on the downstream response
+if a request was dropped due to either :ref:`maintenance mode
+<config_http_filters_router_runtime_maintenance_mode>` or upstream :ref:`circuit breaking
+<arch_overview_circuit_break>`.
+
 .. _config_http_filters_router_stats:
 
 Statistics

docs/configuration/network_filters/network_filters.rst

@@ -14,4 +14,5 @@ filters.
   echo_filter
   mongo_proxy_filter
   rate_limit_filter
+  redis_proxy_filter
   tcp_proxy_filter

docs/configuration/network_filters/redis_proxy_filter.rst

@@ -0,0 +1,97 @@
+.. _config_network_filters_redis_proxy:
+
+Redis proxy
+===========
+
+Redis :ref:`architecture overview <arch_overview_redis>`.
+
+.. code-block:: json
+
+  {
+    "name": "redis_proxy",
+    "config": {
+      "cluster_name": "...",
+      "conn_pool": "{...}",
+      "stat_prefix": "..."
+    }
+  }
+
+cluster_name
+  *(required, string)* Name of cluster from cluster manager.
+  See the :ref:`configuration section <arch_overview_redis_configuration>` of the architecture
+  overview for recommendations on configuring the backing cluster.
+
+conn_pool
+  *(required, object)* Connection pool configuration.
+
+stat_prefix
+  *(required, string)* The prefix to use when emitting :ref:`statistics
+  <config_network_filters_redis_proxy_stats>`.
+
+Connection pool configuration
+-----------------------------
+
+.. code-block:: json
+
+  {
+    "op_timeout_ms": "...",
+  }
+
+op_timeout_ms
+  *(required, integer)* Per-operation timeout in milliseconds. The timer starts when the first
+  command of a pipeline is written to the backend connection. Each response received from Redis
+  resets the timer since it signifies that the next command is being processed by the backend.
+  The only exception to this behavior is when a connection to a backend is not yet established. In
+  that case, the connect timeout on the cluster will govern the timeout until the connection is
+  ready.
+
+.. _config_network_filters_redis_proxy_stats:
+
+Statistics
+----------
+
+Every configured Redis proxy filter has statistics rooted at *redis.<stat_prefix>.* with the
+following statistics:
+
+.. csv-table::
+  :header: Name, Type, Description
+  :widths: 1, 1, 2
+
+  downstream_cx_active, Gauge, Total active connections
+  downstream_cx_protocol_error, Counter, Total protocol errors
+  downstream_cx_rx_bytes_buffered, Gauge, Total received bytes currently buffered
+  downstream_cx_rx_bytes_total, Counter, Total bytes received
+  downstream_cx_total, Counter, Total connections
+  downstream_cx_tx_bytes_buffered, Gauge, Total sent bytes currently buffered
+  downstream_cx_tx_bytes_total, Counter, Total bytes sent
+  downstream_rq_active, Gauge, Total active requests
+  downstream_rq_total, Counter, Total requests
+
+
+Splitter statistics
+-------------------
+
+The Redis filter will gather statistics for the command splitter in the
+*redis.<stat_prefix>.splitter.* with the following statistics:
+
+.. csv-table::
+  :header: Name, Type, Description
+  :widths: 1, 1, 2
+
+  invalid_request, Counter, "Number of requests with an incorrect number of arguments"
+  unsupported_command, Counter, "Number of commands issued which are not recognized by the
+  command splitter"
+
+Per command statistics
+----------------------
+
+The Redis filter will gather statistics for commands in the
+*redis.<stat_prefix>.command.<command>.* namespace.
+
+.. csv-table::
+  :header: Name, Type, Description
+  :widths: 1, 1, 2
+
+  total, Counter, Number of commands
+
+.. _config_network_filters_redis_proxy_per_command_stats:

docs/intro/arch_overview/arch_overview.rst

@@ -28,6 +28,7 @@ Architecture overview
   tcp_proxy
   mongo
   dynamo
+  redis
   hot_restart
   dynamic_configuration
   init

docs/intro/arch_overview/circuit_breaking.rst

@@ -25,10 +25,14 @@ configure and code each application independently. Envoy supports various types
 * **Cluster maximum active retries**: The maximum number of retries that can be outstanding to all
   hosts in a cluster at any given time. In general we recommend aggressively circuit breaking
   retries so that retries for sporadic failures are allowed but the overall retry volume cannot
-  explode and cause large scale cascading failure. If this circuit breaker overflows the 
+  explode and cause large scale cascading failure. If this circuit breaker overflows the
   :ref:`upstream_rq_retry_overflow <config_cluster_manager_cluster_stats>` counter for the cluster
   will increment.
 
 Each circuit breaking limit is :ref:`configurable <config_cluster_manager_cluster_circuit_breakers>`
 and tracked on a per upstream cluster and per priority basis. This allows different components of
 the distributed system to be tuned independently and have different limits.
+
+Note that circuit breaking will cause the :ref:`x-envoy-overloaded
+<config_http_filters_router_x-envoy-overloaded>` header to be set by the router filter in the
+case of HTTP requests.

docs/intro/arch_overview/http_routing.rst

@@ -76,6 +76,9 @@ headers <config_http_filters_router_headers>`. The following configurations are
   requirements. For example, network failure, all 5xx response codes, idempotent 4xx response codes,
   etc.
 
+Note that retries may be disabled depending on the contents of the :ref:`x-envoy-overloaded
+<config_http_filters_router_x-envoy-overloaded>`.
+
 .. _arch_overview_http_routing_priority:
 
 Priority routing