No more than two instances supported for scraping metrics about prometheus_oracle_exporter HOT 3 CLOSED

There is no limitation of instances, but prometheus has a scrape timeout.
Did you manual do a "curl" on the exporter endpoint and looked at the output?
Can you please provide your configuration.

from prometheus_oracle_exporter.

Hi,

The config has this layout (redacted because its a productive system):

connections:
 - connection: user/pwd@instance/SERVICE
   database: xxx
   instance: xxx
   alertlog:
    - file: oracle.conf
#   queries:
#    - sql: "select 1 from dual"
#      name: sample1
#    - sql: "select 2 from dual"
#      name: sample2

 - connection: user/pwd@instance/SERVICE
   database: xzx
   instance: xzx
   alertlog:
    - file: oracle.conf
#   queries:
#    - sql: "select 1 from dual"
#      name: sample1
#    - sql: "select 2 from dual"
#      name: sample2

 - connection: user/pwd@instance/SERVICE
   database: xzy
   instance: xzy
   alertlog:
    - file: oracle.conf
#   queries:
#    - sql: "select 1 from dual"
#      name: sample1
#    - sql: "select 2 from dual"
#      name: sample2

...

I include oracle conf as alert log because I dont have the alertlogs on this host (for how) and the exporter fails if this is missing in the YAML.

Even though there are 8 different host/service entries in my config file only the first two are used for scraping. I added some debug lines during SQL connection establishment and I can see that the exporter connects to all 8 (one of which is down so successfully connects to 7). But only "xxx" and "xyx" in this example are exported.

I did a curl above in the original report.

from prometheus_oracle_exporter.

I've found the solution for this problem. The database and instance names in the config gets overwritten from the instance during scrape. So when they are the same the metrics are overwritten.

From now database and instance gets overwritten when they are empty.
Test 1.1.3

from prometheus_oracle_exporter.