Skip to content

Instantly share code, notes, and snippets.

@mandarjog
Last active October 3, 2018 14:08
Show Gist options
  • Save mandarjog/8a415283f854058f4fe8d309709c3df0 to your computer and use it in GitHub Desktop.
Save mandarjog/8a415283f854058f4fe8d309709c3df0 to your computer and use it in GitHub Desktop.
Standard Istio Experiment result format
# experiment summary
# All times in UTC
start_time: "2018-10-03T12:32:06Z"
end_time: "2018-10-03T12:32:12Z"
# multiple experiments with the same setup should have the same exp_group_uuid
exp_group_uuid: "uuid_string"
istio_build: "build"
# experiments have different standard istio profiles
profile: "full/ingress_only, mixer"
details_uri: "https://ibmcloud-perf.istio.io/regpatrol/release-1.0-20181001-09-15"
env :
kubernetes: gke-1.10.5
node_count: 5
node:
type: "n32"
cores: 32
mem_mb: 4096
config:
istio_policy_enabled: true
istio_telemetry_enabled: true
mtls_enabled: true
proxy_concurrency: 0
client:
# Is the client inside or outside the mesh
# A client inside the mesh will not use ingress_gateway
internal: false
protocol: "http | tcp | grpc"
connections: 20
rps: 1000
# latency histogram in ms and average
latencies_ms:
min: 4
average: 20
p50: 22
p90: 29
p99: 40
max: 45
metrics:
ingress_gateway:
count: 5
# cpu in mCores
cpu_mCores: 2010
# memory in MB
mem_mb: 350
# traffic sent thru ingress gateway
rps: 1000
# Total bytes sent thru ingress
bps: 89000
sidecars:
count: 20
# cpu used by all sidecars except ingress / egress
cpu_mCores: 3000
mem_mb: 600
# Total rps traversing all sidecars
rps: 8000
# Total bytes sent thru sidecars
bps: 213004
istio_telemetry:
count: 20
# cpu used by all istio_telemetry pods
cpu_mCores: 3000
mem_mb: 600
# Total rps traversing all proxies (sidecars+ingress+egress)
rps: 8000
istio_policy:
count: 20
# cpu used by all istio_policy pods
cpu_mCores: 3000
mem_mb: 600
# Total rps traversing all proxies (sidecars+ingress+egress)
rps: 8000
cache_hit_rate: 99
istio_pilot:
count: 20
# cpu used by istio_pilot pods
cpu_mCores: 3000
mem_mb: 600
endpoints: 200
services: 400 # services + service entries
sidecars: 120
virtual_services: 50
destination_rules: 55
# how long does it take a listener change to propagate to 90% proxies
lds_latency_ms: 1020
# how long does it take for cluster change to propagate to 90% proxies
cds_latency_ms: 1020
# additional individual workloads should be listed here
# Only sidecar metrics are measured
individual_workload_1:
name: "fortioserver"
count: 2
cpu_mCores: 55
mem_mb: 100
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment