ConnectWatch — Connector Guide

SSH probe

Monitor Linux/Unix servers by connecting via SSH and writing a test file

Included in all plans Live

What it does — step by step

1

Opens a TCP connection to the server on port 22 and measures connect time

2

Authenticates using your private key or password

3

Opens an SFTP session and uploads a 20-byte test file to the configured path — measures upload time

4

Deletes the test file (no residue left on the server)

5

Records total latency, connect time, upload time. Opens an incident if any step fails.

AWS EC2 / Azure VM

Monitor your cloud VMs. Know immediately if SSH access is lost due to security group changes, reboots, or disk issues.

On-premise servers

Monitor bare metal or VMware servers in your data centre. Detect network outages, power failures, or OS crashes.

Database servers

Confirm that your DB server is reachable over SSH — useful as a baseline health check alongside your DB monitoring.

CI/CD build agents

Ensure your Jenkins or GitHub Actions self-hosted runners are always reachable and writable before builds start.

SSH connect time

Time to establish TCP + SSH handshake. High values indicate network congestion or server load.

Upload time

Time to write the test file via SFTP. High values can indicate disk I/O issues or filesystem problems.

Total latency

End-to-end time from connect to file deleted. Your primary health metric.

Field	Description	Example
hostrequired	Server IP address or hostname	192.168.1.10 or ec2-54-12.compute.amazonaws.com
portoptional	SSH port (default 22)	22
userrequired	SSH username — create a dedicated probe user with minimal permissions	probeuser
keyrequired*	SSH private key (PEM format). Either key or password required.	-----BEGIN OPENSSH PRIVATE KEY-----
passwordrequired*	SSH password (if not using key auth)	••••••••
fs_pathoptional	Path to write the test file. Defaults to /tmp. Used by Filesystem probe too.	/tmp or /mnt/efs/probe

Best practice: Create a dedicated probeuser account with no shell access and write permission only to the probe path. This limits the blast radius if the key is ever compromised.

HTTPS probe

Monitor web services and APIs by sending HTTP requests and verifying responses

Included in all plans Live

What it does — step by step

1

Sends an HTTP GET request to your configured URL

2

Checks the HTTP response code — 2xx is success, anything else is a failure

3

Records response time and HTTP status code

4

Optionally POSTs a small test file to a /probe-upload endpoint if it exists

REST APIs

Monitor your public or internal APIs. Know if your API goes down before your customers report it.

Web applications

Monitor your customer-facing websites. Detect downtime, slow responses, or certificate expiry.

Health endpoints

Poll your /health or /status endpoints to verify the application is running correctly.

Internal services

Monitor internal microservices that aren't publicly accessible but need to be up for your product to function.

Response time

Total time from request sent to response received. Alert if above your SLA threshold.

HTTP status

The response status code. 200 = OK, 5xx = server error, 4xx = client/auth error.

Timeout

If no response within the timeout (default 5s), probe fails. Configurable per environment.

Field	Description	Example
urlrequired	Full URL to probe including protocol	https://api.yourcompany.com/health
headersoptional	JSON object of HTTP headers (auth tokens, API keys)	{"Authorization": "Bearer token"}
timeoutoptional	Request timeout in milliseconds (default 5000)	5000

Tip: For authenticated endpoints, add your API key or Bearer token in the headers field. ConnectWatch encrypts these at rest — they're never stored in plain text.

Filesystem probe

Measure real filesystem write/read latency on NFS, EFS, Azure Files and local disks

Included in all plans Live

What it does — step by step

1

SSHes into the server where the filesystem is mounted

2

Runs dd if=/dev/zero of=/mnt/path/probe.tmp bs=1k count=4 oflag=dsync — the oflag=dsync flag forces a real network flush, bypassing the OS page cache

3

Reads the file back to measure read latency

4

Deletes the test file — no residue left on the filesystem

Why oflag=dsync matters: Without it, the OS caches the write in RAM and reports success immediately — hiding the real NFS/EFS network latency. With dsync, every write is flushed to the network storage before returning, giving you the true end-to-end latency your application experiences.

AWS EFS

EFS has burst credits — when exhausted, throughput drops from 100+ MB/s to 1-5 MB/s. The FS probe catches this degradation before your application slows down.

Azure Files

Azure Files has per-share IOPS limits. Monitor write latency to detect when you're hitting those limits during peak load.

On-premise NFS

NFS mounts can silently degrade due to network issues or NFS server load. The FS probe detects this before users experience slow uploads or saves.

Shared storage

For multi-server setups where all nodes share a filesystem — verify all nodes can write to shared storage independently.

FS write latency

Time to write 4KB and flush to network storage. The most important metric for NFS/EFS health.

FS read latency

Time to read the test file back. High read latency on NFS usually indicates network or NFS server issues.

Total latency

Write + read combined. Compare against your application's file I/O requirements.

Note: The filesystem probe uses SSH to run commands on the server — so it needs its own SSH credentials. These can be the same as your SSH probe or different, since the filesystem may be on a different server.

Full setup guide Server commands, permissions, SSH key setup, error reference & terms

AWS S3

Monitor S3 bucket availability and write/read performance directly via AWS API

$5/env/mo Live

What it does — step by step

1

PUTs a small test object to your S3 bucket using AWS Signature V4 authentication

2

GETs the object back to verify the write landed and measures download time

3

DELETEs the test object — no cost for storage, no residue

Media / file storage

If your app uploads user files to S3, monitor that the bucket is writable and latency is acceptable.

Backups

Verify your backup bucket is accessible before your backup job runs. Catch permission issues early.

Data lake

Monitor the S3 buckets that feed your data pipeline. Detect S3 throttling or outages before they affect analytics.

Static website hosting

For S3-hosted websites, verify the bucket is writable for deployments alongside your HTTPS probe for availability.

Field	Description	Example
bucketrequired	S3 bucket name	my-app-uploads
regionrequired	AWS region where the bucket is located	us-east-1 or ap-south-1
access_key_idrequired	AWS access key ID — create a dedicated IAM user with s3:PutObject, s3:GetObject, s3:DeleteObject only	AKIAIOSFODNN7EXAMPLE
secret_access_keyrequired	AWS secret access key (encrypted at rest)	••••••••
prefixoptional	Object key prefix for test objects	connectwatch-probe/

IAM policy: Create a dedicated IAM user with only these permissions on the specific bucket:
s3:PutObject, s3:GetObject, s3:DeleteObject on arn:aws:s3:::your-bucket/connectwatch-probe/*

Azure Blob Storage

Monitor Azure Blob containers for availability and write/read performance

$5/env/mo Live

What it does

1

Authenticates using Azure Shared Key from your storage account connection string

2

Uploads a BlockBlob test object to your container

3

Downloads it back to verify the write and measure read latency

4

Deletes the test blob

Document storage

Monitor blob containers used for document management, ensuring write availability and acceptable latency.

Media pipelines

Azure Blob is commonly used for video/media ingestion. Monitor that uploads will succeed before your pipeline runs.

Field	Description	Example
connection_stringrequired	Azure Storage account connection string (encrypted at rest)	DefaultEndpointsProtocol=https;AccountName=...
containerrequired	Blob container name	probe-container

SFTP

Monitor dedicated SFTP file transfer services used for business file exchange

$4/env/mo Live

What it does

1

Connects to the SFTP server using key or password authentication

2

Uploads a test file to the configured upload path

3

Reads the file back to verify the upload succeeded

4

Deletes the test file

Bank file exchange

Banks and financial institutions use SFTP to send payment files, statements, and reconciliation data. Monitor that the SFTP endpoint is reachable before your scheduled transfers.

Logistics partners

Shipping and logistics providers use SFTP for order files and tracking data. Detect outages before your order processing pipeline fails.

AWS Transfer Family

Monitor AWS Transfer Family SFTP endpoints — a managed SFTP service commonly used for B2B file exchange.

Government portals

Many government data submission portals use SFTP. Monitor availability before scheduled compliance submissions.

Target

Linux/Unix server (your own infrastructure)

Dedicated SFTP service (yours or a partner's)

Use case

Is my server reachable and healthy?

Is this file transfer service available?

Protocol

SSH + SFTP subsystem

SFTP (may use custom port like 2222)

Auth

Server SSH key (you control it)

SFTP credentials (partner provides them)

Example

Your EC2 instance at 54.12.34.56

Bank's SFTP at sftp.hdfc.com:22

Google Cloud Storage

Monitor GCS buckets for availability and write/read performance

$5/env/mo Live

GCP applications

Monitor GCS buckets used by your Google Cloud applications for file storage, ML datasets, or backups.

ML pipelines

GCS is commonly used for ML training data and model artifacts. Monitor availability before your training jobs start.

Field	Description	Example
bucketrequired	GCS bucket name	my-app-data
service_account_keyrequired	GCP service account JSON key (encrypted at rest). Create a service account with Storage Object Admin role on the bucket only.	{"type":"service_account",...}
prefixoptional	Object prefix for test files	connectwatch-probe/

SharePoint

Monitor SharePoint document libraries for file upload availability

$8/env/mo 🔜 Coming soon

SharePoint connector is in development. It will use the Microsoft Graph API to upload a test file to a document library and verify the upload succeeded.

Planned use cases: Monitor SharePoint Online document libraries used for business file storage, collaboration, and compliance archiving.

SMB / Windows Share

Monitor Windows file shares and NAS devices using the SMB protocol

$4/env/mo 🔜 Coming soon

SMB connector is in development. It will connect to Windows file shares and NAS devices using the SMB/CIFS protocol and write a test file.

Planned use cases: Monitor Windows Server file shares, Synology/QNAP NAS devices, and any CIFS-compatible network storage.

Custom webhook

Push probe results to your own systems for custom alerting and SIEM integration

$10/env/mo 🔜 Coming soon

Custom webhook connector is in development. It will POST probe results in a configurable JSON format to any HTTP endpoint — useful for SIEM tools, custom dashboards, and internal alerting systems.

Storage connector comparison

Choosing the right storage connector for your setup:

Filesystem

NFS mounts, EFS, Azure Files, local disk — anything mounted on a Linux server

SSH key or password to the server

AWS S3

S3 buckets accessed directly via AWS API — not mounted

IAM access key + secret

Azure Blob

Azure Blob containers accessed via Azure Storage API

Storage account connection string

GCS

Google Cloud Storage buckets

GCP service account JSON key

SFTP

Dedicated SFTP services — bank transfers, partner file exchange

SFTP username + key or password

SMB

Windows file shares, NAS devices on local network

Windows domain or local credentials

Quick rule of thumb:
— If it's a Linux server with a mounted filesystem → use Filesystem probe
— If it's an AWS S3 bucket accessed via API → use AWS S3
— If it's a partner's SFTP server for file exchange → use SFTP
— If it's a Windows share or NAS → use SMB (coming soon)

Connector guide

Storage connector comparison