services
The convox services command lists the services defined for an app along with their domains and port mappings. Use convox services restart to restart a specific service without affecting the others.
services
List services for an app
Usage
convox services
Examples
$ convox services
SERVICE DOMAIN PORTS
web web.myapp.0a1b2c3d4e5f.convox.cloud 443:3000
When the app's budget cap has been breached (3.24.6+), convox services
adds a BUDGET column showing the per-service sub-state (armed-Nm,
at-cap-keda, at-cap-auto, at-cap) — see convox ps
for the full value reference and Budget Caps
for recovery flows.
services restart
Restart a service
Usage
convox services restart <service>
Examples
$ convox services restart web
Restarting web... OK
services update
Update a service in place. Mirrors the effect of editing scale.* fields in
convox.yml and redeploying, without a new release. Accepts the same flags as
convox scale (--count, --cpu, --memory, --gpu, --gpu-vendor).
Usage
convox services update <service> [--count N] [--cpu M] [--memory M] [--gpu N] [--gpu-vendor VENDOR]
Examples
$ convox services update web --gpu 1 --gpu-vendor nvidia
Updating web... OK
$ convox services update web --count 3 --memory 2048
Updating web... OK
services triggers enable
Enable a Console-driven autoscaler for a service. Requires rack version 3.24.6 or later. See Autoscale Triggers Override for the full surface (Console + CLI parity).
Usage
convox services triggers enable <service> \
--min <int> --max <int> \
[--cpu <1-100>] [--memory <1-100>] \
[--gpu <1-100>] [--queue <int>]
At least one of --cpu, --memory, --gpu, --queue is required.
--gpu and --queue require KEDA on the rack (keda_enable=true).
--gpu additionally requires the service to declare scale.gpu.count >= 1
in convox.yml.
CPU- or memory-only overrides materialize a native Kubernetes HPA and
require --min >= 1; the Kubernetes HPAScaleToZero feature gate is
alpha and not enabled on managed clusters. For scale-to-zero behavior,
include a KEDA-eligible trigger (--gpu or --queue) — the KEDA
ScaledObject path supports --min 0 natively. Mixed trigger sets
that include any of --gpu / --queue (e.g. --cpu 70 --gpu 75)
dispatch through the KEDA ScaledObject path and therefore also
accept --min 0.
Examples
$ convox services triggers enable web --min 1 --max 5 --cpu 70
Enabling triggers override on web... OK
$ convox services triggers enable worker --min 0 --max 10 --gpu 75 --queue 100
Enabling triggers override on worker... OK
services triggers disable
Remove the Console-driven autoscaler. The next deploy re-materializes the manifest's autoscale config (if any).
Usage
convox services triggers disable <service>
Examples
$ convox services triggers disable web
Disabling triggers override on web... OK
services triggers threshold-set
Update a single trigger's threshold on a service that already has an
override active. --type accepts cpu, memory, gpu, or queue.
Usage
convox services triggers threshold-set <service> \
--type <cpu|memory|gpu|queue> --threshold <number>
Examples
$ convox services triggers threshold-set web --type cpu --threshold 80
Setting web cpu threshold to 80... OK
See Also
- Service for service configuration
- Load Balancers for load balancer setup
- Autoscale Triggers Override for the Console UI behind these CLI subcommands
