# Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root. # Configuration of dispatch from container nodes to search clusters namespace=vespa.config.search ## The active docs a group must have as a % of the average active docs of all other groups, ## for that group to be included in queries minActivedocsPercentage double default=97.0 ## Distribution policy for group selection distributionPolicy enum { ROUNDROBIN, BEST_OF_RANDOM_2, LATENCY_AMORTIZED_OVER_REQUESTS, LATENCY_AMORTIZED_OVER_TIME, ADAPTIVE } default=ADAPTIVE ## Maximum number of hits that will be requested from a single node ## in this dataset. If not set, there is no limit. Using this option ## may help reduce network traffic when searching in datasets with big ## fan-out, but it will also result in incorrect and incomplete results; ## don't use it if you don't (really) mean it. maxHitsPerNode int default=2147483647 ## Probability for getting the K best hits (topK). ## A value of 1.0 will ask all N partitions for K hits. ## Any value between <0, 1> will use a Student T with 30 degrees freedom and compute a value Q that ## will give you the globally K best hits according to this formula with the desired probability. ## q = k/n + qT (p',30) x √(k × (1/n) × (1 − 1/n)) ## With a probability of 0.999 and K=200 and N=10 will give a Q of 38, meaning that you only need to fetch 19% compared to ## a setting of 1.0. This is a significant optimisation with with very little loss in presicion. topKProbability double default=0.9999 ## Number of document replicas _per group_ that will be present in a stable cluster. ## Should always be >= searchableCopies. redundancy long default=1 ## Minimum search coverage required before returning the results of a query minSearchCoverage double default=100 ## Minimum wait time for full coverage after minimum coverage is achieved, factored based on time left at minimum coverage minWaitAfterCoverageFactor double default=0 ## Maximum wait time for full coverage after minimum coverage is achieved, factored based on time left at minimum coverage maxWaitAfterCoverageFactor double default=1 ## Number of JRT transport threads numJrtTransportThreads int default=8 ## Number of JRT connections per backend node numJrtConnectionsPerNode int default=8 ## Number of seconds to spend warming up code to prevent JIT cold start issues. warmuptime double default=0.1 ## Specifies how summary data are decoded ## Eager will build java objects immediately, while ONDEMAND will do so when it is needed summaryDecodePolicy enum {EAGER, ONDEMAND} default=EAGER