[ATLAS-4554] Elasticsearch: Add Elasticsearch backend monitor && Fix some bugs when using atlas with elasticsearch backend

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 3.0.0, 2.3.0
• Component/s: atlas-core, atlas-intg
• Flags: Patch

Language

java

• Assignee: Xingyu Zha
• Reporter: Xingyu Zha
• Created: 14/Feb/22 10:21
• Updated: 09/Oct/22 03:56
• Resolved: 16/Feb/22 19:14

Description

When using elasticsearch as the background index, the status of solr is displayed as unhealthy.

if (!isHealthy && (prevSolrHealthCheckTime == 0 || currentTime - prevSolrHealthCheckTime > SOLR_STATUS_LOG_FREQUENCY_MS)) {
    LOG.info("Solr Health: Unhealthy!"); 

Looking at the code, it is found that atlas currently only has the health monitoring of the solr index, so the relevant monitoring function about elasticsearch has been added.

if (backendSolr.equals(backend) && isSolrHealthy()) {
    isHealthy = true;
    LOG.info("Backend Solr Health: Healthy!");
}
else if (backendElasticSearch.equals(backend) && isElasticsearchHealthy()) {
    isHealthy = true;
    LOG.info("Backend ElasticSearch Health: Healthy!");
} 

By the way, ElasticSearch7index support was added, extracted from the org.janusgraph.diskstorage.es code, and adjusted with method to new client instance

graphdb/janus/src/main/java/org/janusgraph/diskstorage/es/ElasticSearch7Index.java 

At the same time, in the process of use, there is also a problem that cannot be used due to log.info problems, and it has been fixed here

LOG.info("Setting " + SOLR_WAIT_SEARCHER_CONF + " = " + getBoolean(SOLR_WAIT_SEARCHER_CONF)); 

Issue Links

• links to: github pr #160