v0.1.6

Patch Changes

• 22d34a3: Add nipa db inspect, which takes a database cluster's name or ID and answers whether it's healthy and how to reach it. It shows the engine and version, status, health and when Trove last checked it, flavor with its vCPUs and RAM, storage, zone, address with the engine's default port, allowed CIDRs, replicas, whether the general and slow query logs are on, and the 5 newest backups. Its last line says the database is healthy and where to connect, or what needs attention, such as a primary that isn't active or a latest backup that failed. --json prints all of it under database, with every backup, the engine's defaultPort and a problems list of what's wrong and where. nipa db ls --json gives each primary id, zone, vcpus, ramMb, allowedCidrs and healthCheckedAt too, and every status list colors completed backups green and failed ones red.
• 22d34a3: nipa db ls shows a database's internal address when it has no external IP. The Address column was blank for these, and --json gave externalAddress as "" instead of null.
• f31bf3b: Add nipa doctor, which checks your setup in one go and says what to fix: the config files, which profile the run uses and why, whether its Keystone and Space API answer and how fast, the session and when it expires, whether Keystone and the Space API accept its token, the folder's link, openstack and terraform, and a newer nipa from the last update check. It changes nothing and never asks for a password, so it works in scripts. --json prints each check's id, status, summary and hint. It exits with code 1 when a check fails, and a warning doesn't change the exit code.
• 49bc449: Add nipa flavor ls, which lists the flavors a server can have, smallest first, with their vCPUs, RAM and type. The Space portal calls them machine types, so nipa machine-types does the same, and so does nipa flavors. The ones only database clusters use don't show. In a pipe, it prints one flavor name per line.
• 49bc449: Add nipa k8s ls, which lists the Kubernetes clusters in your project with their Kubernetes version, nodes, active node count and age. nipa os coe can't find Magnum on Nipa Cloud production, so nipa finds each cluster through the servers Magnum made for it. nipa kubernetes and nipa coe do the same. nipa server ls --json gives each Kubernetes node its version too.
• 61c0eed: Add nipa lb inspect, which takes a load balancer's name or ID and shows why it doesn't serve. It prints the load balancer's status, health, virtual IP, external IP and flavor, then each listener, backend group and member with its status and health, and marks in red each part that isn't ACTIVE and ONLINE. The last line names what stops it, such as 2 of 3 members are down, or says that every part is healthy. --json prints the whole load balancer with its listeners, backend groups and members. Tab completes load balancer names after nipa lb inspect.
• 49bc449: Add nipa network ls, which lists the networks your project can use with their status, type, zone and age. A VPC network is one of your project's private networks, and an external network is a shared pool of external IPs. nipa networks does the same.
• 3c4252d: Add nipa link [project], like the Vercel CLI's vercel link, which saves the profile and a project in .nipa/project.json. Commands in that folder and the folders below it use them, so a Terraform folder for staging needs no -P staging, and nipa switch in other folders doesn't change it. nipa exchanges the token for the linked project before each command and doesn't save it. nipa unlink removes the link, and nipa whoami --json gives link.
• 51f4c90: Add nipa open, which opens the profile's Space portal in your browser, like vercel open. nipa open server, volume, network, sg, lb or db opens that list, and a name or ID after it opens that resource's page. In a pipe, or with --url, nipa prints only the URL and doesn't open a browser. The portal opens the project you last used in it, because a URL can't pick one.
• 92bda92: Tab completion after nipa os completes server names after server show, stop, start, delete and the other commands that take only servers, and the names after --flavor, --image and --network. nipa reads them from the Space API, keeps them for a minute, and never asks for a password on a Tab press.
• 0cb45a4: Add nipa quota ls, which shows each quota of your project, such as servers, vCPUs, RAM and volumes, by group, with how much it uses, its limit and the percent used. A quota at 80% or more shows in yellow and one at its limit in red, and a note counts them, so you can see why a create failed or how much room is left. --json prints each quota's group, name, used, limit and unit, and a pipe gets one tab-separated line per quota with its unit. nipa quota, nipa quotas and nipa limits do the same.
• fe436ef: Add nipa login --remember, which saves your password in the macOS Keychain, or with secret-tool on Linux, so later logins ask only for an OTP code. Keystone has no refresh token, so a session still lasts 24 hours. When Keystone refuses a saved password, nipa deletes it and asks. nipa logout and nipa profile rm delete the saved password. Commands that use a session with less than 30 minutes left print a note with the login command first.
• 9df4221: Add nipa server history and nipa server logs to find out why a server failed. history, or events, lists the actions on a server, newest first, with who asked, the result and the request ID. logs prints the server's console log, cloud-init output included, and --tail 20 keeps the last 20 lines. A server that never booted has no console log, so nipa says so and names nipa server history. nipa server inspect now shows Nova's task while the server is busy, whether it's locked, and its last action, such as create failed 26m ago by Ann, and names history and logs for a server in Error. --json prints all of it, and inspect --json adds taskState, locked and lastAction to the server.
• 9afaa9b: Add nipa server inspect, start, stop and restart, which take a server's name or ID. inspect shows the server's flavor with its vCPUs and RAM, zone, addresses, volumes and security groups, and --json prints them. start, stop and restart call the Space API, like the Space portal's buttons, then wait until the server finishes. stop and restart ask first, and --yes skips the question. --no-wait returns once the Space API takes the action, and --timeout 10m waits longer than the default 5 minutes. Tab completes server names after these commands. nipa server ls --json gives each server vcpus, ramMb, zone, volumes and securityGroups too.
• 340528d: nipa sg inspect <Tab> completes security group names.
• 340528d: Add nipa sg inspect, which takes a security group's name or ID and shows its description, the servers that use it with their addresses, and its inbound and outbound rules. Each rule shows its protocol, ports, and the CIDR or security group it allows by name. A note says when an inbound rule opens SSH, RDP or a database port to the internet. --json prints the group in the shape nipa sg ls --json uses, and adds each rule's remoteGroupName and the group's servers.
• 49bc449: Add nipa sg ls, which lists the security groups in your project with their inbound and outbound rule counts, age and description. --json prints each rule's direction, protocol, ports and remote. nipa security-group and nipa security-groups do the same.
• 49bc449: Add nipa volume ls, which lists the block storage volumes in your project with their status, size, type, the server each one is attached to, and age. nipa volumes does the same. Volume statuses such as in-use and creating get the same colors as server statuses, and read as "In use" and "Creating".

v0.1.5

Patch Changes

• 197c333: Add nipa completion --install, which saves the tab completion script where bash, zsh or fish loads it, so the shell doesn't run nipa each time it starts. nipa rewrites the saved script when its version changes, so new commands complete after an update. nipa completion without a shell uses the one in $SHELL.
• 47797cd: nipa server ls marks the servers that a Kubernetes cluster made, such as k8s-control-plane-1 (Kubernetes master), from the Magnum tags the portal reads too. --json gives each server kubernetes, which is null for a server you made, or the cluster's clusterId and the node's role. The troubleshooting page explains why nipa os coe and other commands can't find their service on production.
• 587c9ac: Give production's Space API only to profiles on production's Keystone. In 0.1.4, a staging profile got production's Space API, which refused its token, so nipa server ls said the session had expired. nipa now drops that URL from such profiles. nipa profile add --space-url takes the Space portal URL, such as https://portal-stg-epc.nipa.cloud, saves its /api, and checks that it answers. With another Keystone and a terminal, profile add asks for the portal URL. Without one, server ls, db ls, lb ls and ip ls say the profile has no Space API URL before they ask for a login. A profile that 0.1.4 saved with a portal URL gets its /api when nipa loads it, and a Space API URL that answers with a web page gets an error that says so.

v0.1.4

Patch Changes

• 3adea28: Add nipa db ls, which lists the managed database clusters in your project with their engine, status, address, flavor and age. nipa server ls doesn't show them.
• 3adea28: Add nipa ip ls, which lists the external IPs in your project with their status, the internal IP each one forwards to, zone and name. An IP without an internal IP isn't attached to anything.
• 3adea28: Add nipa lb ls, which lists the load balancers in your project with their status, health, virtual IP, listener count and age.
• 49a0027: nipa server ls calls Nipa Cloud's Space API at https://space.nipa.cloud/api on port 443, instead of the compute endpoint in Keystone's catalog on port 8774, which many networks can't reach. A profile keeps its Space API URL in spaceUrl, and nipa profile add --space-url sets it. nipa server ls --json no longer has network on each address, because the Space API doesn't name it. nipa no longer reads Keystone's catalog or saves its endpoints in auth.json.

v0.1.3

Patch Changes

• 6d2d00f: nipa login and nipa switch get a token on Nipa Cloud again. Nipa's gateway drops the connection for a token without a service catalog, so nipa now asks for your project ID on a first login, takes it from --project, or uses the project from your last login.

v0.1.2

Patch Changes

• 15f0ce5: nipa login asks for your OTP code on Nipa Cloud again. Nipa's gateway sends 401 responses with an empty body, so nipa couldn't read the account's MFA rules and printed an empty Error: instead.

v0.1.1

Patch Changes

• 004f686: Install nipa from npm with npm install -g nipa-cli. The package runs on Node.js 22.13 or later, and the binaries on GitHub releases still need nothing else installed.

v0.1.0

Minor Changes

• 93c35d9: nipa profile rm --help and every other subcommand print their own help page, and nipa server --json runs nipa server ls --json. Usage errors name the command and show its usage line, and an extra argument, as in nipa switch Beta extra, exits with code 2.
• 268f3ff: Add --debug, or NIPA_DEBUG=1, to log the method, URL, status and time of each HTTP request.
• ab32643: First release: nipa login with a password and OTP code, logout, whoami, switch, env, exec, os, tf and completion for bash, zsh, fish and PowerShell.
• 268f3ff: nipa login sends your password first and asks for an OTP code only when Keystone asks for one. After a wrong code, it asks for the next one, up to 3 times, without asking for the password again.
• 268f3ff: Add nipa profile to keep a session for more than one Keystone, such as staging, and -P <name> or NIPA_PROFILE to pick one for a command. Files from nipa 0.1 load as the prod profile.
• 503fa77: Add nipa server ls to list the servers in your project with their status, address and flavor, without the OpenStack client.
• 268f3ff: nipa checks GitHub for a newer release at most once a day and prints a notice when there is one. Set NIPA_NO_UPDATE_CHECK to turn it off.
• 268f3ff: nipa shows its prompts, spinners and results in the Vercel CLI style, groups nipa --help by task, and completes global options and the nipa profile subcommands in every shell.

Patch Changes

• 268f3ff: nipa stops reading its own options at --. In nipa switch -- -h, the command gets -h and nipa doesn't print its help.
• 268f3ff: Tab completion after nipa os lists openstack's commands and options in bash, zsh, fish and PowerShell. The OpenStack client ships completion only for bash, so nipa os used to complete file names in the other shells.
• 503fa77: nipa's > messages, such as > Logging in to prod, no longer print in red in a terminal.
• 268f3ff: A mistyped command name gets a suggestion, such as "Did you mean nipa login?".