Add network policy related metadata to flow records #2163

heanlan · 2021-05-11T04:25:18Z

Add fields: network policy type, rule name to flow records of flow exporter
Update related unit-tests, e2e tests
Make necessary changes on elk-flow-collector and flow-visibility doc

codecov-commenter · 2021-05-11T05:46:32Z

Codecov Report

Merging #2163 (2395aa1) into main (d8d8626) will increase coverage by 3.59%.
The diff coverage is 77.27%.

@@            Coverage Diff             @@
##             main    #2163      +/-   ##
==========================================
+ Coverage   61.83%   65.42%   +3.59%     
==========================================
  Files         276      276              
  Lines       21145    21183      +38     
==========================================
+ Hits        13074    13858     +784     
+ Misses       6702     5918     -784     
- Partials     1369     1407      +38

Flag	Coverage Δ
e2e-tests	`56.65% <77.27%> (?)`
kind-e2e-tests	`53.10% <77.27%> (+0.13%)`	⬆️
unit-tests	`41.31% <45.71%> (+0.03%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
pkg/flowaggregator/flowaggregator.go	`62.71% <ø> (+4.66%)`	⬆️
pkg/agent/flowexporter/utils.go	`66.66% <55.55%> (-4.77%)`	⬇️
.../flowexporter/connections/conntrack_connections.go	`78.33% <60.00%> (+1.14%)`	⬆️
pkg/agent/controller/networkpolicy/packetin.go	`72.81% <77.77%> (+0.95%)`	⬆️
pkg/agent/flowexporter/exporter/exporter.go	`78.58% <100.00%> (+1.28%)`	⬆️
...ntroller/networkpolicy/networkpolicy_controller.go	`69.45% <0.00%> (-0.65%)`	⬇️
pkg/ovs/openflow/ofctrl_bridge.go	`50.36% <0.00%> (+0.36%)`	⬆️
pkg/agent/openflow/network_policy.go	`76.41% <0.00%> (+0.59%)`	⬆️
pkg/apiserver/certificate/cacert_controller.go	`58.94% <0.00%> (+0.66%)`	⬆️
... and 47 more

srikartati

Thanks for adding more netpol info in flow records. Could you improve the flow aggregator e2e test as well, where we check the netpol info?

heanlan · 2021-05-11T20:05:36Z

Thanks for adding more netpol info in flow records. Could you improve the flow aggregator e2e test as well, where we check the netpol info?

For sure. I'm also trying to add the e2e test, just facing some interface issue. If I understood correctly, we are using packages networking.v1, meta.v1 to initialize network policies, in flowaggregator_test.go. The interfaces include policy UID, but don’t include policy type, rule name, rule priority(I don't think we need to test on rule priority, since it's actually the ovs flow priority internally computed from tier priority, policy priority, rule priority).

Do we need to change to other interfaces in order to construct network policies including policy type and rule name? I noticed antrea/pkg/apis/controlplane/types.go offers that kind of interface. To make those changes, I might need to change the createNetworkPolicy, deleteNetworkPolicy functions in framework.go, along with other e2e test files which make use of those two functions.

srikartati · 2021-05-11T21:58:14Z

For sure. I'm also trying to add the e2e test, just facing some interface issue. If I understood correctly, we are using packages networking.v1, meta.v1 to initialize network policies, in flowaggregator_test.go. The interfaces include policy UID, but don’t include policy type, rule name, rule priority(I don't think we need to test on rule priority, since it's actually the ovs flow priority internally computed from tier priority, policy priority, rule priority).

Do we need to change to other interfaces in order to construct network policies including policy type and rule name? I noticed antrea/pkg/apis/controlplane/types.go offers that kind of interface. To make those changes, I might need to change the createNetworkPolicy, deleteNetworkPolicy functions in framework.go, along with other e2e test files which make use of those two functions.

Sure, I think those fields policy UID are present in Antrea network policy. Instead of changing the existing test with K8s policy, can you add a new test with the Antrea network policy for one of the flow cases?

zyiou

Don't forget to test with elk flow collector. Some changes are needed as we are adding new fields.

zyiou · 2021-05-12T18:06:13Z

pkg/agent/flowexporter/connections/connections.go

+					conn.IngressNetworkPolicyRuleName = rule.Name
+					// TODO: tier name, tier priority, policy priority to be added when support available
+					if policy.Type == "K8sNetworkPolicy" {
+						conn.IngressNetworkPolicyRulePriority = -1


Better to use the const defined in go-ipfix

zyiou · 2021-05-12T18:12:36Z

pkg/flowaggregator/flowaggregator.go

+		"ingressNetworkPolicyUID",
+		"ingressNetworkPolicyType",
+		"ingressNetworkPolicyRuleName",
+		"ingressNetworkPolicyRulePriority",


I believe go-ipfix has not supported correlation of data type signed32, we should add that before using it here.

Not sure if I get it. I have already changed rule priority's type to signed32 in go-ipfix. In the output of template set, it shows rule priority's length to be 4.

I get it.. Will make more correlated changes in go-ipfix

zyiou · 2021-05-12T18:12:50Z

pkg/flowaggregator/flowaggregator.go

+		"egressNetworkPolicyUID",
+		"egressNetworkPolicyType",
+		"egressNetworkPolicyRuleName",
+		"egressNetworkPolicyRulePriority",


same above.

zyiou · 2021-05-12T18:17:51Z

pkg/agent/flowexporter/connections/connections.go

 				if policy == nil {
 					// This should not happen because the rule flow ID to rule mapping is
 					// preserved for max(5s, flowPollInterval) even after the rule deletion.
 					klog.Warningf("Cannot find NetworkPolicy that has rule with egressOfID %v", egressOfID)
 				} else {
 					conn.EgressNetworkPolicyName = policy.Name
 					conn.EgressNetworkPolicyNamespace = policy.Namespace
+					conn.EgressNetworkPolicyUID = string(policy.UID)
+					switch policy.Type {
+					case "K8sNetworkPolicy":


It would be better to use const from definitions likepkg/apis/controlplane/types.go instead of string here.

zyiou · 2021-05-12T18:18:18Z

pkg/agent/flowexporter/connections/connections.go

+						conn.EgressNetworkPolicyType = 3
+					}
+					conn.EgressNetworkPolicyRuleName = rule.Name
+					if policy.Type == "K8sNetworkPolicy" {


zyiou · 2021-05-12T18:20:24Z

pkg/agent/flowexporter/connections/connections_test.go

+			expConn.EgressNetworkPolicyUID = string(np2.UID)
+			switch np2.Type {
+			case cpv1beta.K8sNetworkPolicy:
+				expConn.EgressNetworkPolicyType = 1


we can use const defined in go-ipfix instead of self-assigned value, same for next several uses.

srikartati

Few comments in the e2e test, otherwise LGTM

srikartati · 2021-05-17T22:33:17Z

test/e2e/flowaggregator_test.go

 		}
 	})

 	// InterNodeFlows tests the case, where Pods are deployed on different Nodes
 	// and their flow information is exported as IPFIX flow records.
 	t.Run("InterNodeFlows", func(t *testing.T) {
-		np1, np2 := deployNetworkPolicies(t, data, "perftest-a", "perftest-c")
+		anp1, anp2 := deployAntreaNetworkPolicies(t, data, "perftest-a", "perftest-c")


A comment that says Antrea network policy is being tested here right before the test and at "IntraNodeFlows" K8s network policy being tested would be useful.

srikartati · 2021-05-17T22:34:02Z

test/e2e/flowaggregator_test.go

@@ -153,33 +172,31 @@ func testHelper(t *testing.T, data *TestData, podAIPs *PodIPs, podBIPs *PodIPs,
 			}
 		}()
 		// TODO: Skipping bandwidth check for Intra-Node flows as it is flaky.
+		// test on K8s network policies


Move the comment to right before the test.
Change the name of deployNetworkPolicies to deployK8sNetworkPolicies and similarly for related functions.

srikartati · 2021-05-17T22:36:09Z

test/e2e/flowaggregator_test.go

 					// Check if records have both ingress and egress network policies.
 					if !strings.Contains(record, ingressNetworkPolicyName) {
 						t.Errorf("Record does not have NetworkPolicy name with ingress rule")
 					}
 					if !strings.Contains(record, fmt.Sprintf("%s: %s", "ingressNetworkPolicyNamespace", testNamespace)) {
 						t.Errorf("Record does not have correct ingressNetworkPolicyNamespace")
 					}
+					if !strings.Contains(record, "ingressNetworkPolicyUID") {


Could we add a custom UID when adding configuring the policy and validate that here?
You have to change the underlying createNetworkPolicy function a little bit.

For K8s network policy, the UID can be customized. But for Antrea network policy, the AntreaNetworkPolicySpecBuilder doesn't support us to customize the UID. Do you think that's ok for us to only validate K8s np's UID? I feel testing on K8s np is indeed good enough, but it may looks a bit inconsistent, for testing on the different fields with different flow types?

Testing only for k8s NP is fine with me.

srikartati · 2021-05-17T22:36:38Z

test/e2e/flowaggregator_test.go

 					if !strings.Contains(record, egressNetworkPolicyName) {
 						t.Errorf("Record does not have NetworkPolicy name with egress rule")
 					}
 					if !strings.Contains(record, fmt.Sprintf("%s: %s", "egressNetworkPolicyNamespace", testNamespace)) {
 						t.Errorf("Record does not have correct egressNetworkPolicyNamespace")
 					}
+					if !strings.Contains(record, "egressNetworkPolicyUID") {


srikartati · 2021-05-17T22:58:53Z

test/e2e/flowaggregator_test.go

+					if !strings.Contains(record, fmt.Sprintf("%s: %s", "ingressNetworkPolicyNamespace", testNamespace)) {
+						t.Errorf("Record does not have correct ingressNetworkPolicyNamespace")
+					}
+					if !strings.Contains(record, "ingressNetworkPolicyUID") {


srikartati · 2021-05-17T22:59:41Z

test/e2e/framework.go

+	busyboxImage  = "projects.registry.vmware.com/library/busybox"
+	nginxImage    = "projects.registry.vmware.com/antrea/nginx"
+	perftoolImage = "projects.registry.vmware.com/antrea/perftool"
+	// ipfixCollectorImage = "projects.registry.vmware.com/antrea/ipfix-collector:v0.5.1"


are you not able to access harbor regisstry?

Have changed it back. Thanks for reminding!

I don't think you pushed?

Just pushed. I was waiting for Srikar's reply on the comments of e2e test. Sorry for the confusion.

antoninbas

sorry for the late review

antoninbas · 2021-05-19T01:09:38Z

test/e2e/framework.go

+	busyboxImage  = "projects.registry.vmware.com/library/busybox"
+	nginxImage    = "projects.registry.vmware.com/antrea/nginx"
+	perftoolImage = "projects.registry.vmware.com/antrea/perftool"
+	// ipfixCollectorImage = "projects.registry.vmware.com/antrea/ipfix-collector:v0.5.1"


I don't think you pushed?

antoninbas · 2021-05-19T01:12:38Z

build/yamls/elk-flow-collector/logstash/ipfix.yml

+  114:
+    - :string
+    - :ingressNetworkPolicyUID
+  115:
+    - :uint8
+    - :ingressNetworkPolicyType
+  116:
+    - :int32
+    - :ingressNetworkPolicyPriority
+  117:
+    - :string
+    - :egressNetworkPolicyUID
+  118:
+    - :uint8
+    - :egressNetworkPolicyType
+  119:
+    - :int32
+    - :egressNetworkPolicyPriority


while adding the policy type makes sense to me, I am not so sure about adding the UID and the priority... The UID is duplicate information once you have the type and the name. The priority is also unnecessary IMO since we can retrieve it based on the rule name. We are just growing the record size, but I don't see much benefit. If a tool needs to retrieve this information, it could be given access to the K8s API. Any opinion on this @jianjuns and @tnqn?

For UID, yes, it means nothing for most users and it cannot be used to get a resource from K8s API. Normally it's only useful to distinguish resources having same name while existing in different time. If the records are only used for flow visibility, I would agree UID has no much value.
For priority, it seems that it's the openflow priority which is dynamically assigned on each Node, so even same policy rule will have different values when agents export them. Will it be really useful?

Regarding UID, the use case we thought of earlier is not to show to the user but it's for integration with systems, where there is access to k8s API of the cluster and UID could be used to fetch the network policy object rather than using the name.
For priority, I was not aware that each agent computes its own priority for a given rule. If that is the case, I agree that it cannot be used. We should show policy priority or tier priority. IMO, we can remove this rule priority for now.

Normally it's only useful to distinguish resources having same name while existing in different time.

@tnqn Do you think it is a case that we need to handle here? If yes, it feels like we should also include the UID for the other K8s resources as well, e.g. source & destination Pods.

I feel like it's all or nothing. Either we include UIDs for all K8s resources, in which case it makes the records larger and adds complexity, or we include it for none of them. The advantage of including the UID is to make sure that we get the correct resource by name from the K8s API.

Regarding UID, the use case we thought of earlier is not to show to the user but it's for integration with systems, where there is access to k8s API of the cluster and UID could be used to fetch the network policy object rather than using the name.

@srikartati But UID in K8s can not be used to fetch object, the primary key in API and etcd is [namespace/]name.

@antoninbas I feel resources with same name usually mean same thing, and pods' names are usually auto-generated randomly so unlikely conflict.. Maybe we can start from not including it and consider adding it when user complains not being able to distinguish records of historial objects causes real problem.

@srikartati But UID in K8s can not be used to fetch object, the primary key in API and etcd is [namespace/]name.

@tnqn I meant in systems where the K8s objects are stored in a separate inventory and given that it is organized by using UID. At the same time, we can request a different inventory organization as well.. not a concrete use-case.

I see Antonin's point that we may need UID for all objects in that case. I agree that we can remove it for now and take it up in the future if there is a strong requirement.

Agreed. Let's remove it for now. It's easier to just add something later if we actually need it.

tnqn · 2021-05-17T15:50:27Z

pkg/agent/flowexporter/connections/connections.go

 	"antrea.io/antrea/pkg/querier"
+	"github.com/vmware/go-ipfix/pkg/registry"


wrong place

tnqn · 2021-05-19T13:18:10Z

pkg/agent/flowexporter/connections/connections.go

@@ -23,12 +23,15 @@ import (
 	corev1 "k8s.io/api/core/v1"
 	"k8s.io/klog/v2"

+	"github.com/vmware/go-ipfix/pkg/registry"


should be in the same group as "k8s.io/klog/v2"

fixed, thanks!

tnqn · 2021-05-19T13:45:45Z

build/yamls/elk-flow-collector/logstash/ipfix.yml

+  114:
+    - :string
+    - :ingressNetworkPolicyUID
+  115:
+    - :uint8
+    - :ingressNetworkPolicyType
+  116:
+    - :int32
+    - :ingressNetworkPolicyPriority
+  117:
+    - :string
+    - :egressNetworkPolicyUID
+  118:
+    - :uint8
+    - :egressNetworkPolicyType
+  119:
+    - :int32
+    - :egressNetworkPolicyPriority


For UID, yes, it means nothing for most users and it cannot be used to get a resource from K8s API. Normally it's only useful to distinguish resources having same name while existing in different time. If the records are only used for flow visibility, I would agree UID has no much value.
For priority, it seems that it's the openflow priority which is dynamically assigned on each Node, so even same policy rule will have different values when agents export them. Will it be really useful?

srikartati · 2021-05-20T17:38:12Z

build/yamls/elk-flow-collector/logstash/ipfix.yml

+  114:
+    - :string
+    - :ingressNetworkPolicyUID
+  115:
+    - :uint8
+    - :ingressNetworkPolicyType
+  116:
+    - :int32
+    - :ingressNetworkPolicyPriority
+  117:
+    - :string
+    - :egressNetworkPolicyUID
+  118:
+    - :uint8
+    - :egressNetworkPolicyType
+  119:
+    - :int32
+    - :egressNetworkPolicyPriority


Regarding UID, the use case we thought of earlier is not to show to the user but it's for integration with systems, where there is access to k8s API of the cluster and UID could be used to fetch the network policy object rather than using the name.
For priority, I was not aware that each agent computes its own priority for a given rule. If that is the case, I agree that it cannot be used. We should show policy priority or tier priority. IMO, we can remove this rule priority for now.

srikartati · 2021-05-20T17:38:55Z

test/e2e/flowaggregator_test.go

 	networkingv1 "k8s.io/api/networking/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/util/wait"
+
+	crdv1alpha1 "antrea.io/antrea/pkg/apis/crd/v1alpha1"
+	. "antrea.io/antrea/test/e2e/utils"


do you need a period here?

removed, thanks

srikartati · 2021-05-20T17:40:16Z

test/e2e/flowaggregator_test.go

 					// Check if records have both ingress and egress network policies.
 					if !strings.Contains(record, ingressNetworkPolicyName) {
 						t.Errorf("Record does not have NetworkPolicy name with ingress rule")
 					}
 					if !strings.Contains(record, fmt.Sprintf("%s: %s", "ingressNetworkPolicyNamespace", testNamespace)) {
 						t.Errorf("Record does not have correct ingressNetworkPolicyNamespace")
 					}
+					if !strings.Contains(record, "ingressNetworkPolicyUID") {


Testing only for k8s NP is fine with me.

zyiou · 2021-05-20T17:45:28Z

docs/network-flow-visibility.md

+| ingressNetworkPolicyType         | 56506         | 115      | unsigned8   |
+| ingressNetworkPolicyRuleName     | 56506         | 111      | string      |
+| ingressNetworkPolicyRulePriority | 56506         | 116      | signed32    |
+| egressNetworkPolicyName          | 56506         | 141      | string      |


zyiou · 2021-05-20T17:46:02Z

docs/network-flow-visibility.md

+| ingressNetworkPolicyNamespace    | 56506         | 111      | string      |
+| ingressNetworkPolicyUID          | 56506         | 114      | string      |
+| ingressNetworkPolicyType         | 56506         | 115      | unsigned8   |
+| ingressNetworkPolicyRuleName     | 56506         | 111      | string      |


Is the field ID correct?

fixed, thanks for reminding

zyiou · 2021-05-20T17:49:48Z

test/e2e/flowaggregator_test.go

+	ingressNetworkPolicyUID: ac42f0ba-f580-46ae-8336-2bfc45024a9a
+	ingressNetworkPolicyType: 2
+	ingressNetworkPolicyRuleName: test-ingress-rule-name
+	ingressNetworkPolicyRulePriority: 50000


extra indent here

have no idea why there is an extra indent, which doesn't exist locally.
There is also extra indent in the main:

antrea/test/e2e/flowaggregator_test.go

Line 70 in c21592c

flowType: 1

zyiou · 2021-05-20T17:49:57Z

test/e2e/flowaggregator_test.go

+	egressNetworkPolicyUID: bbbc0320-b461-45e3-987a-d2d7caf7aad3
+	egressNetworkPolicyType: 2
+	egressNetworkPolicyRuleName: test-egress-rule-name
+	egressNetworkPolicyRulePriority: 50000
 	flowType: 1


heanlan · 2021-05-27T16:10:40Z

/test-all

heanlan · 2021-05-27T23:10:19Z

/test-all

heanlan · 2021-05-28T01:10:26Z

/test-e2e

heanlan · 2021-05-28T01:46:17Z

/test-networkpolicy

pkg/agent/flowexporter/connections/conntrack_connections.go

tnqn

LGTM

antoninbas · 2021-05-28T02:44:10Z

/test-all

change to const in unit-test add e2e test add new fields to doc changes on elk-flow-collector side change dependency module and image name change to const change ipfix dependency back to before v0.5.1 fix api issue Signed-off-by: heanlan <[email protected]>

heanlan · 2021-05-28T02:45:40Z

/test-all

heanlan · 2021-05-28T05:32:58Z

jenkins-e2e failed on somewhere unrelated with this PR. TestFlowAggregator passed all the tests.

--- PASS: TestFlowAggregator (225.18s)
--- PASS: TestFlowAggregator/IPv4 (86.92s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeFlows (14.68s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeDenyConnIngressANP (3.14s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeDenyConnEgressANP (3.67s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeDenyConnNP (5.09s)
--- PASS: TestFlowAggregator/IPv4/InterNodeFlows (13.77s)
--- PASS: TestFlowAggregator/IPv4/InterNodeDenyConnIngressANP (3.19s)
--- PASS: TestFlowAggregator/IPv4/InterNodeDenyConnEgressANP (3.11s)
--- PASS: TestFlowAggregator/IPv4/InterNodeDenyConnNP (9.14s)
--- PASS: TestFlowAggregator/IPv4/LocalServiceAccess (14.34s)
--- PASS: TestFlowAggregator/IPv4/RemoteServiceAccess (13.68s)

srikartati · 2021-05-28T05:32:59Z

Flow aggregator tests have passed:
--- PASS: TestFlowAggregator (225.18s)
--- PASS: TestFlowAggregator/IPv4 (86.92s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeFlows (14.68s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeDenyConnIngressANP (3.14s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeDenyConnEgressANP (3.67s)
--- PASS: TestFlowAggregator/IPv4/IntraNodeDenyConnNP (5.09s)
--- PASS: TestFlowAggregator/IPv4/InterNodeFlows (13.77s)
--- PASS: TestFlowAggregator/IPv4/InterNodeDenyConnIngressANP (3.19s)
--- PASS: TestFlowAggregator/IPv4/InterNodeDenyConnEgressANP (3.11s)
--- PASS: TestFlowAggregator/IPv4/InterNodeDenyConnNP (9.14s)
--- PASS: TestFlowAggregator/IPv4/LocalServiceAccess (14.34s)
--- PASS: TestFlowAggregator/IPv4/RemoteServiceAccess (13.68s)

TestUserProvidedCert test has failed.
@tnqn I am running the test again. Could you please monitor this PR and merge?
/test-e2e

tnqn · 2021-05-28T05:58:09Z

@tnqn I am running the test again. Could you please monitor this PR and merge?

@srikartati Sure!

vmwclabot added the cla-not-required label May 11, 2021

heanlan force-pushed the flow-records-metadata branch 2 times, most recently from 4ae17e4 to 9bc8e51 Compare May 11, 2021 19:00

srikartati reviewed May 11, 2021

View reviewed changes

srikartati added this to the Antrea v1.1 release milestone May 12, 2021

zyiou reviewed May 12, 2021

View reviewed changes

heanlan marked this pull request as draft May 12, 2021 20:25

heanlan force-pushed the flow-records-metadata branch 3 times, most recently from 831ef4a to 0b02e77 Compare May 14, 2021 22:20

heanlan marked this pull request as ready for review May 14, 2021 22:34

heanlan marked this pull request as draft May 14, 2021 22:47

heanlan marked this pull request as ready for review May 14, 2021 22:58

heanlan marked this pull request as draft May 14, 2021 23:24

heanlan force-pushed the flow-records-metadata branch from 5dfd494 to f99bac8 Compare May 15, 2021 06:37

heanlan marked this pull request as ready for review May 15, 2021 06:47

heanlan requested review from srikartati and zyiou May 17, 2021 19:59

heanlan force-pushed the flow-records-metadata branch 2 times, most recently from cadc529 to 19d8b58 Compare May 17, 2021 21:54

srikartati reviewed May 17, 2021

View reviewed changes

antoninbas reviewed May 19, 2021

View reviewed changes

heanlan force-pushed the flow-records-metadata branch from 19d8b58 to 02c5a6f Compare May 19, 2021 03:12

tnqn reviewed May 19, 2021

View reviewed changes

heanlan force-pushed the flow-records-metadata branch from 02c5a6f to 15320b0 Compare May 20, 2021 01:53

srikartati reviewed May 20, 2021

View reviewed changes

zyiou reviewed May 20, 2021

View reviewed changes

heanlan dismissed antoninbas’s stale review via 3702994 May 27, 2021 16:10

heanlan force-pushed the flow-records-metadata branch from b632244 to 3702994 Compare May 27, 2021 16:10

heanlan force-pushed the flow-records-metadata branch 2 times, most recently from 42aaf8e to 63a951b Compare May 27, 2021 23:09

heanlan requested a review from antoninbas May 27, 2021 23:10

antoninbas previously approved these changes May 27, 2021

View reviewed changes

tnqn reviewed May 28, 2021

View reviewed changes

pkg/agent/flowexporter/connections/conntrack_connections.go Show resolved Hide resolved

heanlan dismissed antoninbas’s stale review via bc91393 May 28, 2021 02:38

heanlan force-pushed the flow-records-metadata branch from 63a951b to bc91393 Compare May 28, 2021 02:38

tnqn previously approved these changes May 28, 2021

View reviewed changes

add metadata to flow records

2395aa1

change to const in unit-test add e2e test add new fields to doc changes on elk-flow-collector side change dependency module and image name change to const change ipfix dependency back to before v0.5.1 fix api issue Signed-off-by: heanlan <[email protected]>

heanlan dismissed tnqn’s stale review via 2395aa1 May 28, 2021 02:44

heanlan force-pushed the flow-records-metadata branch from bc91393 to 2395aa1 Compare May 28, 2021 02:44

tnqn approved these changes May 28, 2021

View reviewed changes

srikartati mentioned this pull request May 28, 2021

Fix bug in FlowAggregator when sending template set #2039

Merged

antoninbas merged commit 3560475 into antrea-io:main May 28, 2021

heanlan deleted the flow-records-metadata branch May 28, 2021 07:50

		"antrea.io/antrea/pkg/querier"
		"github.com/vmware/go-ipfix/pkg/registry"

Add network policy related metadata to flow records #2163

Add network policy related metadata to flow records #2163

Conversation

heanlan commented May 11, 2021 • edited Loading

codecov-commenter commented May 11, 2021 • edited Loading

Codecov Report

srikartati left a comment

Choose a reason for hiding this comment

heanlan commented May 11, 2021 • edited Loading

srikartati commented May 11, 2021

zyiou left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

srikartati left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

antoninbas left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

heanlan May 24, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

heanlan commented May 27, 2021

heanlan commented May 27, 2021

heanlan commented May 28, 2021

heanlan commented May 28, 2021

tnqn left a comment

Choose a reason for hiding this comment

antoninbas commented May 28, 2021

heanlan commented May 28, 2021

heanlan commented May 28, 2021 • edited Loading

srikartati commented May 28, 2021 • edited Loading

tnqn commented May 28, 2021

heanlan commented May 11, 2021 •

edited

Loading

codecov-commenter commented May 11, 2021 •

edited

Loading

heanlan commented May 11, 2021 •

edited

Loading

heanlan May 24, 2021 •

edited

Loading

heanlan commented May 28, 2021 •

edited

Loading

srikartati commented May 28, 2021 •

edited

Loading