diff options
author | Arne H Juul <arnej27959@users.noreply.github.com> | 2021-12-15 18:36:25 +0100 |
---|---|---|
committer | GitHub <noreply@github.com> | 2021-12-15 18:36:25 +0100 |
commit | 0f6bbcc4d5b29397c7c9b6842a7ebc3749be1850 (patch) | |
tree | 096c6b0671dee2cd2e3313c4f02a24bd7a418504 | |
parent | 1829af6e8a3c23bafd9a13ce5ee3296847c7d069 (diff) | |
parent | 4bd2d6ebb745ee1f90630a44741383f038bde835 (diff) |
Merge pull request #20512 from vespa-engine/arnej/ignore-some-user-agents
minimal implementation of "ignore some user agents"
5 files changed, 39 insertions, 7 deletions
diff --git a/config-model/src/main/java/com/yahoo/vespa/model/container/http/JettyHttpServer.java b/config-model/src/main/java/com/yahoo/vespa/model/container/http/JettyHttpServer.java index 8c31eabfde0..d84884665ef 100644 --- a/config-model/src/main/java/com/yahoo/vespa/model/container/http/JettyHttpServer.java +++ b/config-model/src/main/java/com/yahoo/vespa/model/container/http/JettyHttpServer.java @@ -23,6 +23,7 @@ public class JettyHttpServer extends SimpleComponent implements ServerConfig.Pro private final ContainerCluster<?> cluster; private volatile boolean isHostedVespa; private final List<ConnectorFactory> connectorFactories = new ArrayList<>(); + private final List<String> ignoredUserAgentsList = new ArrayList<>(); public JettyHttpServer(String componentId, ContainerCluster<?> cluster, boolean isHostedVespa) { super(new ComponentModel(componentId, com.yahoo.jdisc.http.server.jetty.JettyHttpServer.class.getName(), null)); @@ -44,10 +45,15 @@ public class JettyHttpServer extends SimpleComponent implements ServerConfig.Pro return Collections.unmodifiableList(connectorFactories); } + public void addIgnoredUserAgent(String userAgent) { + ignoredUserAgentsList.add(userAgent); + } + @Override public void getConfig(ServerConfig.Builder builder) { builder.metric(new ServerConfig.Metric.Builder() .monitoringHandlerPaths(List.of("/state/v1", "/status.html", "/metrics/v2")) + .ignoredUserAgents(ignoredUserAgentsList) .searchHandlerPaths(List.of("/search")) ); if (isHostedVespa) { diff --git a/container-core/abi-spec.json b/container-core/abi-spec.json index a6783d1e5f5..c3c946ab537 100644 --- a/container-core/abi-spec.json +++ b/container-core/abi-spec.json @@ -1990,11 +1990,14 @@ "public com.yahoo.jdisc.http.ServerConfig$Metric$Builder monitoringHandlerPaths(java.util.Collection)", "public com.yahoo.jdisc.http.ServerConfig$Metric$Builder searchHandlerPaths(java.lang.String)", "public com.yahoo.jdisc.http.ServerConfig$Metric$Builder searchHandlerPaths(java.util.Collection)", + "public com.yahoo.jdisc.http.ServerConfig$Metric$Builder ignoredUserAgents(java.lang.String)", + "public com.yahoo.jdisc.http.ServerConfig$Metric$Builder ignoredUserAgents(java.util.Collection)", "public com.yahoo.jdisc.http.ServerConfig$Metric build()" ], "fields": [ "public java.util.List monitoringHandlerPaths", - "public java.util.List searchHandlerPaths" + "public java.util.List searchHandlerPaths", + "public java.util.List ignoredUserAgents" ] }, "com.yahoo.jdisc.http.ServerConfig$Metric": { @@ -2009,7 +2012,9 @@ "public java.util.List monitoringHandlerPaths()", "public java.lang.String monitoringHandlerPaths(int)", "public java.util.List searchHandlerPaths()", - "public java.lang.String searchHandlerPaths(int)" + "public java.lang.String searchHandlerPaths(int)", + "public java.util.List ignoredUserAgents()", + "public java.lang.String ignoredUserAgents(int)" ], "fields": [] }, diff --git a/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/HttpResponseStatisticsCollector.java b/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/HttpResponseStatisticsCollector.java index 0aa2820f959..631f4080c7e 100644 --- a/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/HttpResponseStatisticsCollector.java +++ b/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/HttpResponseStatisticsCollector.java @@ -3,6 +3,7 @@ package com.yahoo.jdisc.http.server.jetty; import com.yahoo.jdisc.Metric; import com.yahoo.jdisc.http.HttpRequest; +import org.eclipse.jetty.http.HttpHeader; import org.eclipse.jetty.http.HttpStatus; import org.eclipse.jetty.server.AsyncContextEvent; import org.eclipse.jetty.server.Handler; @@ -21,6 +22,7 @@ import java.io.IOException; import java.util.ArrayList; import java.util.Collection; import java.util.HashMap; +import java.util.HashSet; import java.util.List; import java.util.Map; import java.util.Objects; @@ -106,12 +108,25 @@ class HttpResponseStatisticsCollector extends HandlerWrapper implements Graceful } } - private void observeEndOfRequest(Request request, HttpServletResponse flushableResponse) throws IOException { - var metrics = StatusCodeMetric.of(request, monitoringHandlerPaths, searchHandlerPaths); - metrics.forEach(metric -> - statistics.computeIfAbsent(metric, __ -> new LongAdder()) - .increment()); + void ignoreUserAgent(String agentName) { + ignoredUserAgents.add(agentName); + } + private Set<String> ignoredUserAgents = new HashSet<>(); + + private boolean shouldLogMetricsFor(Request request) { + String agent = request.getHeader(HttpHeader.USER_AGENT.toString()); + if (agent == null) return true; + return ! ignoredUserAgents.contains(agent); + } + + private void observeEndOfRequest(Request request, HttpServletResponse flushableResponse) throws IOException { + if (shouldLogMetricsFor(request)) { + var metrics = StatusCodeMetric.of(request, monitoringHandlerPaths, searchHandlerPaths); + metrics.forEach(metric -> + statistics.computeIfAbsent(metric, __ -> new LongAdder()) + .increment()); + } long live = inFlight.decrementAndGet(); FutureCallback shutdownCb = shutdown.get(); if (shutdownCb != null) { diff --git a/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/JettyHttpServer.java b/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/JettyHttpServer.java index fca34f3bbd7..006d6cc84da 100644 --- a/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/JettyHttpServer.java +++ b/container-core/src/main/java/com/yahoo/jdisc/http/server/jetty/JettyHttpServer.java @@ -159,6 +159,9 @@ public class JettyHttpServer extends AbstractServerProvider { new HttpResponseStatisticsCollector(serverConfig.metric().monitoringHandlerPaths(), serverConfig.metric().searchHandlerPaths()); statisticsCollector.setHandler(gzipHandler); + for (String agent : serverConfig.metric().ignoredUserAgents()) { + statisticsCollector.ignoreUserAgent(agent); + } StatisticsHandler statisticsHandler = newStatisticsHandler(); statisticsHandler.setHandler(statisticsCollector); diff --git a/container-core/src/main/resources/configdefinitions/jdisc.http.jdisc.http.server.def b/container-core/src/main/resources/configdefinitions/jdisc.http.jdisc.http.server.def index db7322c76f2..7e1404bbc5e 100644 --- a/container-core/src/main/resources/configdefinitions/jdisc.http.jdisc.http.server.def +++ b/container-core/src/main/resources/configdefinitions/jdisc.http.jdisc.http.server.def @@ -57,6 +57,9 @@ metric.monitoringHandlerPaths[] string # Paths that should be reported with search dimensions where applicable metric.searchHandlerPaths[] string +# User-agent names to ignore wrt statistics (crawlers etc) +metric.ignoredUserAgents[] string + # HTTP request headers that contain remote address accessLog.remoteAddressHeaders[] string |