Frequently Asked Questions

How much does Forge cost?

Forge has three plans:

  • Free: up to 3 servers, 7-day metric retention, community support.
  • Pro ($3/node/month, first 3 nodes free): unlimited servers, 90-day metric retention, unlimited notification channels, AI analysis (Gemma 4), MCP API access, email support. Example: 10 servers is $21/month.
  • Enterprise (contact us): unlimited servers, 1-year metric retention, SSO/SAML, priority support, SLA, and on-premises deployment option.

All plans include the full set of 38 alert rules and the Crucible agent at no extra cost. Pricing is per account, not per server (within plan limits).

Which operating systems does Crucible support?

Crucible runs on Linux. The following distributions are tested and supported:

  • Ubuntu 20.04, 22.04, 24.04
  • Debian 11, 12
  • RHEL 8, 9
  • Rocky Linux 8, 9
  • AlmaLinux 8, 9
  • Arch Linux (rolling)
  • Amazon Linux 2, 2023

Crucible requires systemd and a Linux kernel 4.18 or newer. It works on both x86_64 and aarch64 (ARM64) architectures.

Windows and macOS are not supported. For monitoring non-Linux systems, consider using the Forge API directly with a custom metrics collector.

How much CPU and memory does Crucible use?

Crucible is designed to be lightweight:

  • Memory: approximately 90 MB RSS. Varies by hardware: servers with more IPMI sensors use more.
  • CPU: less than 0.5% of a single core on a typical server. Collection runs are brief (under 500 ms) and occur once per interval (default: 300 seconds).
  • Disk: the binary is about 12 MB. Log files are rotated at 50 MB by default.
  • Network: each metrics push is approximately 2-5 KB of compressed JSON. At the default 300-second interval, this is about 1 KB/minute or 1.5 MB/day.

On machines with many disks (50+) or network interfaces (20+), resource usage may be slightly higher. You can reduce overhead by excluding unused interfaces and disks in the configuration.

Where is my data stored?

Metric data is stored on Glassmkr infrastructure in the EU (Frankfurt, Germany). All data is encrypted at rest using AES-256 and in transit using TLS 1.3.

Forge does not store raw system logs, process lists, or file contents. The data collected is limited to numerical metrics (CPU percentages, memory amounts, disk usage, network counters) and hardware status identifiers (SMART health, RAID state, sensor readings).

You can delete all data for a server at any time by removing the server from the Forge dashboard. Account deletion removes all data within 30 days.

Can I self-host Forge?

Yes, on the Enterprise plan. The self-hosted version runs as a set of Docker containers (API server, database, and a small web frontend). It requires:

  • A Linux server with Docker and Docker Compose.
  • PostgreSQL 15+ (can be the included container or an external instance).
  • At least 2 GB RAM and 20 GB disk for the Forge services.

Self-hosted Forge is functionally identical to the SaaS version. Updates are delivered as new container images. Contact [email protected] for pricing and setup assistance.

How do I uninstall Crucible?

To completely remove Crucible from a server:

# Stop and disable the service
sudo systemctl stop crucible
sudo systemctl disable crucible

# Remove the binary
sudo rm /usr/local/bin/crucible

# Remove the systemd unit file
sudo rm /etc/systemd/system/crucible.service
sudo systemctl daemon-reload

# Remove configuration and logs
sudo rm -rf /etc/glassmkr
sudo rm -rf /var/log/glassmkr

Optionally, delete the server from the Forge dashboard to remove its stored metrics and free up a server slot on your plan.

Does Crucible need root access?

The Crucible agent runs as root by default because reading SMART data, IPMI sensors, and certain system files requires elevated privileges. If you do not need SMART, IPMI, or ECC monitoring, you can run Crucible as a non-root user by disabling those collectors and ensuring the user has read access to /proc and /sys.

How long is metric data retained?

Retention depends on your plan:

  • Free: 7 days at full resolution.
  • Pro: 90 days. Data older than 7 days is downsampled to 5-minute resolution. Data older than 30 days is downsampled to 1-hour resolution.
  • Enterprise: 1 year. Same downsampling schedule as Pro for data older than 30 days.

Can I monitor Docker containers or Kubernetes pods?

Crucible currently monitors the host system, not individual containers. Container-level CPU and memory usage are visible in the host metrics as part of the total. Dedicated container and Kubernetes monitoring is on the roadmap.

If you run Crucible inside a Docker container, it needs access to the host's /proc, /sys, and device files via volume mounts. This is not the recommended setup; installing directly on the host is simpler and more reliable.

Can I export my data?

Yes. You can pull metric data via the Health History API endpoint. The response is JSON and can be piped into any analytics tool. For bulk exports, contact support for a CSV or Parquet dump of your account's data.

Does Forge support two-factor authentication?

Yes. Go to Settings > Account > Security and enable TOTP-based two-factor authentication. Forge supports any TOTP app (Google Authenticator, Authy, 1Password, etc.). Recovery codes are provided during setup.

What is per-core CPU monitoring?

When per_core: true is set in the CPU collector configuration (requires Crucible 0.3.0+), Crucible reports individual CPU core utilization in addition to aggregate metrics. This enables the per-core CPU chart in the dashboard expanded view and gives the AI analysis engine awareness of per-core load patterns. This is useful for detecting single-threaded bottlenecks, core pinning issues, and uneven load distribution. Per-core monitoring increases data volume proportionally to the core count.

How does alert muting work?

You can mute specific alert rules on a per-server basis from the server detail page or via the configuration file. Muted rules are not evaluated and do not fire notifications. This is useful during maintenance windows, RAID rebuilds, or when a known condition is expected. Unmuting takes effect on the next ingest cycle. Alerts do not fire retroactively for conditions that occurred while muted.

What do the P1-P4 priority levels mean?

Every alert has a priority from P1 (critical, immediate action required) to P4 (informational). Priority determines the badge color on alert cards and the emoji prefix in Telegram/Slack notifications. P1 alerts indicate data loss or service outage. P2 alerts indicate significant degradation. P3 alerts are early warnings. P4 alerts are proactive recommendations. See the Alert Priorities documentation for the full mapping.

How does AI analysis work?

On the Pro plan, Forge uses Gemma 4 to analyze your server health data when alerts fire. The AI reviews the current metrics (including per-core CPU data if available), recent trends, and the alert context to provide a one-sentence summary of the likely cause. AI analysis is shown on the alert card in the dashboard and included in Telegram/Slack notifications. It is a diagnostic aid, not a replacement for human investigation.

What happens if the Forge API is unreachable?

Crucible continues collecting metrics locally and buffers up to 60 data points (about 5 hours at the default 300-second interval) in memory. When connectivity is restored, the buffered data is pushed in order, so you will not lose data during brief outages. If the buffer fills, the oldest data points are dropped first.

How do I update Crucible?

Run the install script again. It detects the existing installation and upgrades the binary in place:

curl -fsSL https://forge.glassmkr.com/install | sudo bash

The service is restarted automatically. Your configuration file is preserved. You can check the current version with:

crucible --version

Is there an API rate limit?

Yes. Authentication endpoints are limited to 10 requests/minute. The ingest endpoint allows 120 requests/minute per server. All other endpoints allow 60 requests/minute per token. See the API Reference for details.

How do I contact support?

Email [email protected] with your account email and server ID. For faster resolution, attach the output of crucible debug, which generates a diagnostic bundle with redacted configuration, logs, and system information.