AS216394 · WireShield

Documentation

Setup, connection and day-to-day operations guides.

Monitoring, notifications, reports and API

10.1 Appliance monitoring (Appliance Admin)

PageContents
Monitoring → Overview (Device Overview)Device front panel (ports, link lights, BYPASS state), device status, hostname, model, platform, serial number, device ID, firmware version, uptime; CPU (average of all cores and the busiest core), RAM, temperature, fans, power usage; Performance Trend (CPU/RAM, last hour); Hardware Health; License & Support; Interfaces (live RX/TX); Update Status. The refresh interval can be chosen (5 s, 15 s, 1 min).
Monitoring → HardwareProcessor (per physical CPU average usage, cores/threads, temperature; expand for cores), memory and storage, sensors (fans, power supplies, power usage, board temperatures from the BMC/IPMI), device identity (vendor, product, serial number, BIOS). A virtual machine shows that no sensors are available.
Monitoring → Ports & Wire PairsInterfaces, link state, XDP mode and counters (5 s); link speed, FEC and RX/TX ring settings, queue count (Chapter 3.5).
System → SystemHost, WireShield daemon (version, start time, data interface), sampling rate and ring loss, audit chain, TLS certificate, interfaces and Drop reasons (counters).
System → DiagnosticsNetwork tools, System reports (CPU cores, network port, connection tracking, host firewall, routes, interfaces, system log, boot; Chapter 13.6) and packet capture (10.11).
System → Audit LogHash-chained, tamper-evident record of operations (last 200): time, user, action, target, source address, result.

The device card at the bottom left shows the hostname and Healthy / Attention.

Health warnings

An attack is not an alarm. Attacks are tracked through attack records, notifications and the bell icon; unless there is a hardware or capacity problem, the device status does not switch to Attention during an attack. Health warnings concern the appliance itself:

WarningMeaning / action
A CPU core has been above 90% utilization for 5 minutesSustained load (a capacity problem). Short spikes during an attack do not warn. See Chapter 13.6.
Fan / Power supply / Temperature sensor reports warning, critical or faultThe BMC's (iLO/iDRAC) own threshold state. Check the server's hardware management interface.
IOMMU translation is on for <port>; boot with iommu=ptThe data port runs in IOMMU translation mode; packet processing costs more CPU. See Chapter 13.6.
No license is installed: traffic through the appliance is limited to 100 MbpsImport the licence (Chapter 11.4).
Memory usage above 90 %, free disk space below 10 %Resource problem; contact support.
The TLS certificate expires soon / has expiredChapter 5.
Protected interface is downCheck the cable, SFP and far end (Chapter 13.6).
No DNS server is configuredName resolution fails (certificates, GeoIP, feeds). Set the servers on Network Settings → Access & TLS → DNS servers (Chapter 3.8).
The audit log is offManagement changes are rejected; contact support.

Appliance Admin → Hardware — physical CPU groups, sensors, BMC fan/PSU information

Appliance Admin → Hardware — physical CPU groups, sensors, BMC fan/PSU information System → Audit Log — record table

System → Audit Log — record table

10.2 Traffic and attack monitoring (MSSP and customer)

  • Overview: in the MSSP view Network Protection Overview (active tenants, total prefixes, active GRE tunnels and BGP sessions, quota alerts, recent attacks, customers and quotas); in the customer view Protection Overview. Charts show a 95th percentile line.
  • Attacks / My Attacks: detected attack records (newest first) with All, Active, Ended filters. Attacks with a significant share of spoofed sources are tagged "Spoofed sources".
  • Protection Groups / My Prefixes: prefixes, mode, live incoming and dropped traffic; click a row for the prefix detail (traffic chart: live 1 s, last 1 h, last 6 h; most targeted hosts; sources; attacks on this prefix).
  • Mitigation awaiting approval (admin/operator): while an attack on a prefix with operator approval waits for a decision, a card with Approve / Deny and a countdown appears at the top of Overview and Attacks (Chapter 7.10).
  • The Search... box in the top bar searches prefixes and attacks. The bell icon shows the number of active attacks. Attack start/end, mode changes and system alerts appear as pop-up notifications.
  • On every time chart you can drag with the mouse to zoom.

MSSP → Overview (Network Protection Overview) — KPI cards, traffic chart, recent attacks

MSSP → Overview (Network Protection Overview) — KPI cards, traffic chart, recent attacks MSSP → Attacks — attack list with Active/Ended filters

MSSP → Attacks — attack list with Active/Ended filters

10.3 Attack detail

Clicking an attack row opens:

  • The header with the main target IP and vector; prefix, start, duration.
  • Peak traffic, start rate, Dropped packets (in learning mode "would drop"), Detection latency (detection → mitigation).
  • Attack traffic chart: live while the attack runs, 10 s resolution, from 5 min before to 5 min after; incoming / dropped / would drop.
  • Traffic to the target IP, Mitigations applied (drop reasons), Vector composition (traffic class), Attack sources (top) (estimated from sampling; with country flags and AS names), target IPs and ports, source country and ASN.
  • Details: event ID, start action, end, vectors, ended by, Approval state. While a decision is pending, admin and operator see Approve and Deny on the Approval line (Chapter 7.10).
  • Operator notes (staff only): type in Add a note… and press Add note.

Actions:

  • End manually: the attack record is closed and the prefix returns to monitoring. If traffic is still under attack it will be detected again. For a false alarm (e.g. the customer's traffic grew), tick This was normal traffic: learn the current level as normal in the confirmation; otherwise the same level is reported again after 2 minutes.
  • Create PDF report (for an ended attack) / PDF report (download) / Regenerate (e.g. with the current version after an update).

During an attack, header samples (60 s before to 60 s after) are kept in the attack archive; inspect them in the console shell with incident samples <id> and incident flows <id>.

Attack detail — header, KPIs, live attack traffic chart, mitigations applied, sources

Attack detail — header, KPIs, live attack traffic chart, mitigations applied, sources "End manually" confirmation — "This was normal traffic" option

"End manually" confirmation — "This was normal traffic" option

10.4 Traffic analysis

MSSP → Monitoring → Traffic Analyzer (customer: Traffic Analysis): choose the Scope (all prefixes or one prefix). The page shows the estimated rate, estimated distinct sources, top source AS / IPs / countries / networks (/24 · /48), target ports, verdicts / drop reasons, TCP flags and packet sizes. A time chart breaks traffic down by country, ASN, protocol, port or target IP (Last 1 h, Last 24 h). Values are estimated from sampling; history is kept in memory and starts again after the service restarts.

Traffic Analyzer — colour time chart by country/ASN and distribution tables

Traffic Analyzer — colour time chart by country/ASN and distribution tables

10.5 Notification channels and rules

MSSP → Management → Integrations. Admin and operator can make changes.

Adding a channel:

  1. Add channel → choose a type:
TypeFields
E-mail (SMTP)SMTP server (host:port), Security (STARTTLS / TLS / None (unencrypted)), Username, SMTP password, From, Recipients (comma separated), Subject prefix, E-mail language, Panel address for buttons (optional), Open to customers (10.6)
WebhookURL (https://…), Format (JSON / Slack), Allow plain http://, Message language, Signing secret (HMAC, optional)
TelegramBot token, Chat ID, message language
SyslogServer — the transport is the address prefix: udp://192.0.2.50:514, tcp://192.0.2.50:514 or tls://192.0.2.50:6514; Format (CEF / JSON); Facility (e.g. local0)
SNMP trapTrap receiver (host:port), Community, Enterprise OID (optional)

With Customer (optional) set, the channel receives only that customer's (non-staff) events. Passwords and secrets are never shown after saving ("saved — leave empty to keep").

  1. Save, then test the channel with Send test.

Adding a rule: channels receive nothing until a rule sends events to them.

  1. Add rule → Name (e.g. "Attacks to NOC"), Events (none selected = all events), channels, Minimum severity, Only this customer (optional), Cooldown (s).
  2. Save.

Quick start: Send attacks and reports to all channels creates a ready rule.

Events and severities:

EventSeverity
Attack started (attack.start)high
Attack ended, with summary (attack.end)info
Mitigation awaiting approval (mitigation.pending)high
Mitigation decision (mitigation.decided)notice
Prefix mode changed (prefix.mode)notice
Protection settings changed (protection.changed)notice
Operator note added (incident.note)—
Report ready (report.ready; PDF attached for e-mail)info
System alert (system.alert)warning / critical

Note: attack.end has severity info, so a rule with Minimum severity = notice does not receive attack-end notifications; write a separate rule for them.

Delivery: failed deliveries are retried after 1 s, 5 s, 30 s, 2 min, 10 min. At most 60 messages per minute per channel; the excess is suppressed and mentioned in the next message. Delivery history (last 30) shows time, channel, event, title, result, attempt and error. For SMS, connect an SMS gateway through a webhook.

MSSP → Integrations — channels table, rules, delivery history

MSSP → Integrations — channels table, rules, delivery history Add channel dialog — E-mail (SMTP) fields

Add channel dialog — E-mail (SMTP) fields

10.6 Customer self-service notifications

E-mail (MSSP side): to open an SMTP channel to customers, tick Open to customers in the channel form and enter the portal address (e.g. https://portal.example.com) under Customer portal address for buttons in customer e-mails. Customers never see the SMTP settings.

Customer side: in the customer portal, Reporting → Notifications:

  • E-mail notifications card: Recipients (one per line, at most 10), E-mail language, Events (Attack started, Attack ended, Report ready, Protection mode changed), Send e-mails switch, Save, Send test e-mail (once per minute). System events are never sent to customers.
  • Webhook card: Address (https://, public internet), Format (JSON (signed) or Slack incoming webhook), Message language, events → Create webhook. The Signing secret is generated by the system and shown only once; if lost, use Rotate signing secret. The address must resolve to the public internet (private/reserved addresses are refused); redirects are not followed.
  • Channels set up by your provider and Recent deliveries are listed.

Customer Portal → Notifications — E-mail notifications and Webhook cards

Customer Portal → Notifications — E-mail notifications and Webhook cards

10.7 Syslog / CEF

A syslog channel sends RFC 5424 messages over UDP, TCP or TLS. The body is ArcSight CEF (default) or the webhook JSON. Example header:

CEF:0|WireShield|DDoS Appliance|<version>|attack_start|<title>|8|...
FieldContents
Signature IDEvent type (attack_start, attack_end, prefix_mode, system_alert, …)
Severity1 info, 3 notice, 5 warning, 8 high, 10 critical
rt, externalId, msgEvent time (ms), event id, title
cs1–cs5vector, prefix, tenant, action, user who made the change
cn1–cn3peak pps, peak bps, duration (s) (attack end)

Facility defaults to local0, app name to wireshield. A private CA can be set for TLS.

10.8 Verifying webhook signatures

Each request carries: Content-Type: application/json, User-Agent: WireShield/<version>, X-WireShield-Timestamp: <unix>, X-WireShield-Signature: sha256=<hex>.

Signature = hex(HMAC-SHA256(secret, timestamp + "." + raw_body)). On the receiver:

  1. Compute it over the raw request body.
  2. Compare with "sha256=" + hex using a constant-time comparison.
  3. Reject timestamps more than 300 s from now.

Body fields: schema, id, event, severity, time, host, version, prefix_id, prefix, tenant, title, body, fields (vector, action, pps, bps, target, incident; values are strings). Any 2xx response is success. The Slack format sends {"text": "title\nbody"}.

10.9 API, event stream and Prometheus

  • API documentation: the appliance serves separate references for customers and MSSP staff: /docs/api/customer.html and /docs/api/mssp.html (the API documentation link on the API tokens page). They include guides for tokens, errors, the event stream, webhooks, syslog and Prometheus.
  • Authentication: Authorization: Bearer wsk_… (Chapter 9.8). Example:
curl -H "Authorization: Bearer $WS_TOKEN" https://panel.example.com:8443/v1/prefixes
  • Live event stream (SSE): GET /v1/events. Events: attack.start, attack.end, prefix.mode, mitigation.decided (both audiences); for staff also mitigation.pending, protection.changed, incident.note, report.ready, system.alert, bgp.session.up/down, bgp.prefix, tunnel.health, escalation and others. On reconnect, Last-Event-ID replays the last 1000 events; a : ping arrives every 15 s. At most 5 streams per user.
  • Prometheus: GET /metrics (on the management listener; not available through the portal). Use a read-only token restricted to the Prometheus server's address:
- job_name: wireshield
  scheme: https
  metrics_path: /metrics
  authorization:
    type: Bearer
    credentials_file: /etc/prometheus/wireshield.token
  static_configs:
    - targets: ['192.0.2.10:8443']

Metrics: data-plane counters and drop reasons, per-prefix rates and mode, attack state, notification delivery, telemetry, audit chain, system CPU/memory/disk, interfaces, licence state.

  • System endpoints (new in 1.10.0): admin only, accepted only on the management interface (not through the portal), every change written to the audit log:
  • Backups (Chapter 11.3): GET /v1/system/backups (list, schedule, remote targets), POST /v1/system/backups (back up now), GET /v1/system/backups/download?name=<file>, POST /v1/system/backups/upload?name=<file> (upload a .wsbak for checking or restoring), POST /v1/system/backups/verify, POST /v1/system/backups/restore, PUT /v1/system/backups/remote (SFTP/S3 settings), POST /v1/system/backups/remote/test, POST /v1/system/backups/remote/push (send the newest backup now).
  • DNS (Chapter 3.8): GET / PUT /v1/system/dns.
  • Rate limits: 20 requests/s per user (burst 40), 60 requests/s per client address.

10.10 Reports

MSSP → Management → Reports (customer: Reporting → My Reports): generated PDF reports (attack, monthly summary, SLA) are listed and downloadable. Customers see only their own reports.

Automatic attack reports:

  1. On the Reports page, switch on Automatic attack reports: a PDF is created automatically after every attack ends (when the attack's traffic record is complete, about 6 minutes later).
  2. Optionally choose an e-mail channel under Also send by e-mail to (add an e-mail channel under Integrations first).

Report types:

TypeContents
Attack reportTraffic chart (incoming/clean/dropped, 10 s), duration, peak bps/pps, dropped data, vectors, response timeline, source IP/ASN/country tables, mitigations applied
Monthly summaryAttack count, total duration, peaks, vectors, top targets
Customer SLAAttack count, average/maximum detection → mitigation time, share mitigated within target, dropped traffic, usage p95 and quota

PDFs are in English. Operator names and notes are not included. Reports created by hand are not e-mailed through notification rules; scheduled and attack-end reports are.

Scheduled reports (admin/operator): the Scheduled reports card lists the schedules with type, customer, When, Next run and Last run; Run now creates the report immediately, Edit and Delete change or remove a schedule (reports already generated are kept).

  1. Press New schedule.
  2. Name and Type: Monthly summary or Customer SLA.
  3. Customer: all customers (monthly summary only) or one customer. For Customer SLA choose the customer and set SLA target: mitigation within (ms) (default 10000).
  4. Repeat: Monthly (with Day of month, 1–28), Weekly (with Day), Daily or Manual only; for the first three also Hour (0–23) and Time zone (e.g. Europe/Istanbul; the panel suggests your browser's time zone). A monthly run covers the previous calendar month, a weekly run the previous 7 days, a daily run the previous day.
  5. Send by e-mail to: tick e-mail channels (only channels of all customers, or of the chosen customer, are offered; add an e-mail channel under Integrations first).
  6. Keep Enabled ticked and press Create.

A monthly summary on the 1st of every month is a good start.

Generate now creates a single report without a schedule: choose Type, Customer, Month (and the SLA target for an SLA report), then Generate. Such reports are not e-mailed.

In the console shell, report list lists the reports and report schedule list the schedules.

MSSP → Reports — report list, the "Automatic attack reports" switch and the "Scheduled reports" card

MSSP → Reports — report list, the "Automatic attack reports" switch and the "Scheduled reports" card

10.11 Packet capture (PCAP)

System → Diagnostics → New capture (admin/operator only):

  1. Choose Prefix, Duration (s, max 300), Protocol, Source IP/CIDR, Destination port.
  2. Start capture.
  3. The Captures table shows state, packets and size; download with PCAP, or Stop / Delete.

Captures contain payload (up to 256 bytes per packet); every download is audited. One capture per prefix, 8 at a time.

Diagnostics → New capture form and Captures table

Diagnostics → New capture form and Captures table

10.12 IPFIX / sFlow export

Flow records built from the appliance's traffic samples can be sent to external collectors (NetFlow/IPFIX analyser, sFlow collector, SIEM). Admins set them up in MSSP → Integrations → Flow export (IPFIX / sFlow):

  1. Click Add collector.
  2. Enter a Name (lowercase letters, digits, -, _), the Type (IPFIX or sFlow v5) and the Collector address (host:port), e.g. 192.0.2.50:4739 for IPFIX or 192.0.2.51:6343 for sFlow.
  3. IPFIX: choose the Transport (UDP or TCP) and optionally the Active flow timeout (5–300 s, default 30). sFlow: optionally the Agent address (this appliance as the collector sees it).
  4. Optionally limit what is sent: Only these prefix IDs and the Send check boxes (passed, dropped, would drop (monitor), challenged (SYN validation)); none ticked = all traffic.
  5. Save. The change applies immediately; unchanged collectors keep running.

The table shows records and bytes sent, the last send time and errors per collector. Exported flows are customer traffic data: send them only to systems you control. In the console shell, export show lists the collectors.

Still have a question about this guide?
Ask our technical team