When testing IPFIX for output sampling for protocol MPLS on core interface, we see MPLS route record lookup failure errors increment continuously, In the customer lab for the PTX10K box running 22.4R3-S3.3-EVO and in Juniper lab we see the same issue.
user@ptx10k-re0> show services accounting errors inline-jflow fpc-slot 0
Sep 30 18:27:22
Error information
FPC Slot: 0
Flow Creation Failures: 0
Route Record Lookup Failures: 17807127, AS Lookup Failures: 9117
Export Packet Failures: 13
Memory Overload: 0, Memory Alloc Fail Count: 0
IPv4:
IPv4 Flow Creation Failures: 0
IPv4 Route Record Lookup Failures: 0, IPv4 AS Lookup Failures: 2370
IPv6:
IPv6 Flow Creation Failures: 0
IPv6 Route Record Lookup Failures: 13565, IPv6 AS Lookup Failures: 2157
MPLS:
MPLS Flow Creation Failures: 0
MPLS Route Record Lookup Failures: 17793562, MPLS AS Lookup Failures: 4590 <<<==incremental
{master}
Sep 30 18:28:11
Route Record Lookup Failures: 18409067, AS Lookup Failures: 9117
MPLS Route Record Lookup Failures: 18395502, MPLS AS Lookup Failures: 4590 <<<==incremental
user@ptx10k-re0>
Currently MPLS Egress sampling is supporting on PTX10k Device as the mpls egress firewall Filters are getting hit and uplifting the sampled packets to Jflow. Jflow is able to process these Sampled packets. In some cases of MPLS Egress sampling, The mpls label which is present in sampled Flow packet does not have a route entry in Global Forwarding table itself, hence Jflow will not be having the same entry in it's route database. Because of this reason, it is leading to Route lookup failure for the particular label every time, which is an expected behaviour.