Age | Commit message (Collapse) | Author | Files | Lines |
|
In previous implementation performed healthcheck on container
was counted as client connection, because metric was notified
for every TCP channel opened. This was making this metric less useful
(to avoid saying "useless" ;) ).
- refactored NettyTcpServer trying to extract functions with logic
not so strictly related to HV-VES behavior. This also resolves
discussions in https://gerrit.onap.org/r/#/c/76274/
- some renames and methods splitting was made in attempt to make code
more readable
- hv-ves should not count connections from either "127.0.0.1" or
"localhost" to his `totalConnections` metric
- removed redundant logging by adding new methods to Logger
Change-Id: I5f10dac8dac82eafd3b0de92a7ec43f2c23b8c16
Issue-ID: DCAEGEN2-1119
Signed-off-by: Filip Krzywka <filip.krzywka@nokia.com>
|
|
Change-Id: I96183e2358ac10457568d223f37d69c61dfd252d
Issue-ID: DCAEGEN2-1117
Signed-off-by: kjaniak <kornel.janiak@nokia.com>
|
|
Change-Id: I34beb32a7c53da97c6945ec8d0022ac37059b7c5
Issue-ID: DCAEGEN2-670
Signed-off-by: Piotr Jaszczyk <piotr.jaszczyk@nokia.com>
|
|
Check if logger invocations don't use unoptimal invocations, eg.
concatenation `debug("a=" + a)` instead of lambda use `debug {"a=" + a}`
Unfortunately to avoid defining dependencies in many places and having
circural dependencies it was necessarry to reorganize the maven module
structure. The goal was to have `sources` module with production code and
`build` module with build-time tooling (detekt rules among them).
Issue-ID: DCAEGEN2-1002
Change-Id: I36e677b98972aaae6905d722597cbce5e863d201
Signed-off-by: Piotr Jaszczyk <piotr.jaszczyk@nokia.com>
|